How technologies are driving digital innovation in mosquito surveillance systems: a global scoping review over the past decade

Abstract

Objectives

This review investigates the recent advancements aimed at optimising vector control through enhanced mosquito monitoring. Recent advancements in artificial intelligence (AI), internet of things (IoT) and remote sensing have been explored to enhance surveillance but a comprehensive review of their progress was needed.

Design

Scoping review following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 guidelines.

Data sources

Systematic searches were performed in PubMed and Scopus focusing on studies published between January 2014 and June 2024

Eligibility criteria

Original studies focusing on mosquitoes and addressing the design, development, testing, implementation, as well as the evaluation of smart traps or modification, upgrade or digitisation of traditional mosquito traps or surveillance systems were included.

Data extraction and synthesis

Key information relevant to this review was extracted into an Excel spreadsheet and narratively synthesised. We assessed the scientific quality of the studies included in this review based on their designs and especially on their resulting outcomes.

Results

This work highlights the increasing integration of AI in mosquito identification and classification, complemented by environmental sensors for real-time monitoring of parameters such as temperature, humidity and landscape features. Additionally, the convergence of AI with IoT systems, robotic traps and remote sensing technologies, including drones and geographic information systems, enables the collection of granular data on mosquito populations and their habitats, facilitating proactive control measures. Overall, these smart surveillance technologies offer a transformative advancement in vector control, enabling more precise and efficient interventions. However, their widespread adoption will require thorough evaluation of cost-effectiveness, scalability and adaptability across diverse settings.

Conclusions

This review maps the evolution of smart mosquito surveillance over the past decade, highlighting AI-driven species identification, IoT-enabled real-time monitoring and the integration of remote sensing technologies. It provides a structured analysis of how these innovations improve mosquito detection and data accuracy. These insights can guide further research, inform practical implementations and shape policy decisions in public health and mosquito surveillance.

WHAT IS ALREADY KNOWN ON THIS TOPIC.

WHAT THIS STUDY ADDS

• This review maps the evolution of smart mosquito surveillance over the past decade, highlighting AI-driven species identification, IoT-enabled real-time monitoring and the integration of remote sensing technologies. It provides a structured analysis of how these innovations improve mosquito detection and data accuracy.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The findings advocate for a shift towards automated, data-driven vector control strategies, enabling more timely and precise interventions. These insights can guide further research, inform practical implementations and shape policy decisions in public health and mosquito surveillance.

Introduction

Mosquito-borne diseases, including malaria, dengue, Zika virus, yellow fever, West Nile virus and Chikungunya, remain significant global public health challenges.¹ According to the WHO, vector-borne diseases account for over 17% of all infectious diseases, leading to more than 700 000 deaths annually. Malaria alone is responsible for over 608 000 deaths each year, with children under five being the most affected. Dengue has experienced a dramatic surge in cases in recent years, placing nearly half of the world’s population at risk. Moreover, other viral diseases transmitted by mosquitoes, such as chikungunya, yellow fever and West Nile virus, further contribute to the increasing global health burden.² Rapid, unplanned urbanisation, changes in land use and the rise in international travel and trade are intensifying human–vector interactions. Additionally, climate change and environmental shifts are accelerating the spread of these diseases worldwide. In recent years, vector-borne diseases have expanded into previously unaffected regions, with many pathogens once confined to tropical and subtropical zones now emerging in temperate areas. This geographical expansion has resulted in persistent illnesses and outbreaks across diverse regions.³

Furthermore, the economic burden of vector-borne diseases is considerable, affecting both healthcare expenditures and overall economic productivity. Research indicates that malaria, in particular, has been linked to reduced economic growth rates. Between 1965 and 1990, countries with endemic malaria experienced annual per capita economic growth rates 0.25% to 1.3% lower than those free from the disease.^{4 5} In 2015, global investments in malaria control and elimination reached approximately US$2.9 billion.⁶ The financial toll of other vector-borne diseases is similarly significant. For example, in 2013, the worldwide cost of managing dengue was estimated at US$8.9 billion.⁷ Likewise, the economic impact of the Zika virus outbreak has been estimated at roughly US$8.9 billion globally.⁸

Traditionally, mosquito surveillance has relied on deploying multiple traps across target areas to capture live mosquitoes, which are then transported to laboratories for identification and analysis. However, these conventional traps have inherent limitations in real-world conditions, as they frequently capture a wide range of insects, not just disease-carrying mosquitoes. This lack of specificity necessitates labor-intensive and time-consuming manual sorting in laboratories, which slows down the response time in vector control efforts.⁹ To overcome these limitations, modern surveillance systems have integrated advanced technologies, including geographic information systems (GIS), automated mosquito identification systems and internet of things (IoT) platforms enhanced by machine learning algorithms. These innovations significantly improve data accuracy and enable more sophisticated analyses, thereby supporting informed decision-making in vector control strategies.¹⁰ Furthermore, they further enhance the planning and implementation of vector control interventions, optimising resource allocation and reducing the time required for surveillance, ultimately enhancing the effectiveness of interventions.⁶

Despite growing interest in smart mosquito surveillance systems, the landscape of these technologies is rapidly evolving, with new methods and devices being continuously introduced. Despite the potential benefits, a comprehensive synthesis of existing evidence is needed to understand the scope, implementation and effectiveness of these smart systems. This scoping review aims to fill this gap by mapping out the current state of research, identifying and critically evaluating key trends, challenges and opportunities in the field of smart mosquito surveillance. This approach is well suited to capture the breadth and complexity of this emerging field, supporting improved decision-making and intervention strategies.

The primary objective of this scoping review is to systematically map the existing literature on the digital transformation of mosquito surveillance systems. The review highlights the specific electronic and data-driven technologies that characterise these ‘smart systems’, their applications and their impact on mosquito control and public health outcomes. Specifically, it aims to identify the broad spectrum of smart technologies employed in mosquito surveillance, encompassing electronic hardware (eg, sensors, IoT devices and robotics) and software components (eg, artificial intelligence (AI), machine learning and computer vision), in accordance with the established search strategy (see table 1). Although other technological fields such as molecular methods (eg, eDNA), advances in chemical attractants or purely statistical modelling are also progressing, the scope of this review is delimited to integrated digital, hardware and software innovations. Furthermore, the review seeks to evaluate the effectiveness of these systems by synthesising evidence on their capacity to enhance mosquito monitoring, improve data accuracy and contribute to vector control efforts and public health outcomes. Finally, it examines potential opportunities for future research and development within this digital domain.

Table 1. Search terms used in the literature search.

Category	Search terms
Mosquito Genera	mosquitoes; Anopheles gambiae; Anopheles stephensi; Anopheles; Aedes; Aedes aegypti; Aedes albopictus; Culex; Culex quinquefasciatus
Mosquito surveillance methods	smart trap; mosquito trap; mosquito surveillance; vector control; malaria control; mosquito monitoring; vector surveillance; entomology; mosquito sampling; mosquito classification; mosquito control; mosquito tracking; insect tracking
Technologies	optical sensor; machine learning; digital holography; metrology; infrared video tracking; three dimensional localization; insecticide treated bed net; trajectory analysis; retro-reflective screen; imaging; algorithm; artificial intelligence; unmanned aerial vehicle; uav; uncrewed aerial vehicle; drone; deep learning; remote sensing; stratification; IoT; real time environmental monitoring; ovitrap; prediction model; computer vision; GIS; infrared sensors; convolutional neural network; CNN; diffuse reflectance spectroscopy; AI model; robot

Search terms listed within each category were combined using the ‘OR’ operator. The three main categories (‘mosquito genera’, ‘mosquito surveillance methods’ and ‘advanced technologies’) were then combined using the ‘AND’ operator to form the final query. These search terms were derived from both a preliminary literature survey and our expertise and knowledge in the field. Some less advanced but widespread technologies that are often combined with newer technologies have been included allowing us to get a widened search scope and include articles on technologies that may have been outside our knowledge.

Materials and methods

To provide a comprehensive overview of recent technological advancements in mosquito surveillance, we conducted a literature search using both PubMed and Scopus with similar queries. The literature search was independently conducted by four authors, restricting the search to the last decade from 1 January 2014 to 19 June 2024.

Research protocol

The review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Extension for Scoping Reviews guidelines.¹¹ A protocol was developed and agreed on by all authors to systematically identify and analyse relevant studies on mosquito surveillance and control using advanced technologies. A comprehensive search was performed in PubMed and Scopus, focusing on three main categories: mosquito genera, mosquito control methods and advanced technologies. The keywords for each category, connected through OR operators are detailed in table 1. These categories were then combined using the AND operator to ensure that retrieved studies addressed all three aspects. The detailed search strings used for each database are provided in online supplemental file 1. The ‘mosquito genera’ terms were strategically selected to focus on the most epidemiologically significant mosquito vectors that have shown the greatest expansion of their geographic range in recent years. Authoritative reviews identify the most widespread and invasive species as belonging exclusively to the Anopheles, Aedes and Culex genera.¹² Genera outside this scope, such as Mansonia, were therefore not included. The logic of the search categories detailed in table 1 was to align with the specific focus of this review on integrated smart systems. The ‘technologies’ category was designed to capture the core digital, electronic and data-driven components (eg, AI, IoT and optical sensor). Among this category, we have expressly included some less advanced but widespread technologies that are often combined with newer technologies (for instance, insecticide-treated bed nets, while an older approach, are frequently combined with other, more advanced collection technologies), which allowed us to widen the search scope. The ‘mosquito surveillance methods’ terms were used to situate these technologies within practical applications. Therefore, while the search was not designed to capture innovations in fields such as chemistry or statistics in isolation, studies on these topics were included when they were explicitly integrated with a smart sensor, AI or IoT component, reflecting the review’s focus on digital and electronic surveillance systems. Moreover, stratification methods are not a technology themselves, even though they classify geographic areas according to environmental, epidemiological, or vector-risk factors. However, these methods were included because mosquito surveillance programs commonly use them to enhance risk mapping and guide the application of advanced tools such as GIS, remote sensing, and AI-based modeling.

The search was restricted to English-language studies published between 2014 and 2024, focusing on recent technological advancements. The full query was designed to capture a wide range of studies on the implementation and evaluation of smart traps and smart surveillance systems.

Identified studies underwent a two-step screening process: first, titles and abstracts were reviewed, followed by full-text assessments, based on specific eligibility criteria. Data were extracted on study characteristics, methodologies, technologies used and key findings, and the results were synthesised to provide a comprehensive overview of recent advancements in mosquito surveillance and control technologies.

Eligibility criteria

The eligibility criteria, summarised in table 2, were defined according to the Population, Intervention, Comparison, Outcome framework¹² to ensure a structured and transparent study selection process.

Table 2. Eligibility criteria for sources of evidence according to Population, Intervention, Comparison, Outcome (PICO) framework.

PICO	Include	Exclude
Population	Mosquitoes	Studies not involving mosquito population
Intervention/Exposure	Research that addresses the design, development, testing, implementation or evaluation of smart traps or modification/upgrade/digitisation of traditional mosquito traps or mosquito surveillance systems	Research that does not address the design, development, testing, implementation or evaluation of smart traps or modification/upgrade/digitisation of traditional mosquito traps or mosquito surveillance systems
Comparator/Context	Traditional mosquito surveillance system without advanced enhancements
Outcome	Effectiveness in mosquito surveillance (eg, accuracy in mosquito detection, species identification) and control (eg, reduction in mosquito populations, prevention of mosquito-borne diseases)	Studies that do not report relevant outcomes related to mosquito surveillance and control
Study characteristics	Original journal articles, conference papers and technical/scientific reports	Letters to editors, editorials, reviews, scoping reviews, meta-analyses, news articles and media reports
Other	Language: English Time frame: 2014–2024	Languages other than English Time frame: studies before 2014

• Population: studies focusing on mosquito populations were included; studies not involving mosquitoes or mosquito-related surveillance were excluded.

• Intervention: we included studies addressing the design, development, testing, implementation or evaluation of smart mosquito surveillance technologies. This encompassed smart traps as well as the modification, upgrade or digitisation of traditional mosquito traps or surveillance systems. Both standalone smart systems and hybrid approaches combining conventional and advanced technologies were considered. Studies not addressing mosquito surveillance interventions were excluded.

• Comparator: no specific comparator was required for inclusion. When applicable, traditional mosquito surveillance systems without advanced technological enhancements were considered as an implicit baseline for contextual comparison.

• Outcomes: to ensure that included technologies had been applied within a mosquito surveillance or control context, studies were required to report at least one surveillance-related or control-related outcome. Outcomes of interest included, but were not limited to, effectiveness of mosquito surveillance (eg, detection performance, species identification accuracy, monitoring capability) and control-related indicators (eg, mosquito population reduction, support for prevention of mosquito-borne diseases). Studies presenting only conceptual, descriptive or purely technical developments without any application to mosquito surveillance or control were excluded. This outcome requirement was adopted to distinguish implemented or tested surveillance technologies from purely conceptual or preapplication proposals, while still allowing inclusion of early-stage or pilot studies reporting minimal or preliminary performance indicators.

The review included original journal articles, conference papers, and technical or scientific reports. Letters to editors, editorials, reviews, systematic reviews, meta-analyses, news articles and media reports were excluded. Only studies published in English between 2014 and 2024 were considered, while those published before 2014 were excluded.

Study selection and data extraction

The selection process involved an initial screening of titles and abstracts to identify potentially relevant studies. This was followed by a full-text review of selected studies to confirm they met the inclusion criteria. Any disagreements at each stage were resolved through discussions among the reviewers. To ensure consistency and accuracy in data extraction, a scoping approach was employed to chart data from the included sources of evidence. A data charting form was specifically developed to capture relevant information from each study. These forms were designed based on the key research questions and objectives of the scoping review, and included sections for:

• Study identification (study ID, title, country study was conducted, and settings, ie, type of environment the study was conducted).

• Aim of the study.

• Methodology (study design and summary of the methodology used).

• Population (mosquito species targeted and total sample size).

• Areas of intervention (categorised into five areas: AI applications, AI for environmental applications, sensor-based technologies for environmental applications, AI and sensors, and sensor technologies, where sensor refers to studies that use sensor technologies/robots/IoT and environment includes studies that employ environmental monitoring and remote sensing of environmental variables; in addition, models or systems used in the studies were also included when present).

• Outcomes (primary outcome and additional outcomes when available including their means and unit of measurement).

• Results (quantitative results and statistical analysis performed).

The areas of intervention were categorised based on the prevalent themes observed in the studies reviewed. It is important to note that these categories are not exclusively independent, and one study can fall into multiple categories. This flexible categorisation allowed us to capture the multifaceted nature of modern mosquito surveillance and control strategies, reflecting the integrated approaches used in literature. Before their implementation, the data charting forms were pre-tested by the review team to ensure clarity, comprehensiveness and usability. This pretesting phase involved using the forms on a sample of studies to identify any ambiguities or gaps in the data collection process. Adjustments were made to the forms based on the feedback received during this phase.

To minimise bias and enhance the reliability of the data charting process, the reviewers independently charted data from each included study. This approach ensured that each study was evaluated separately by all reviewers, reducing the likelihood of errors and discrepancies. After the independent charting, the reviewers compared their extracted data. Any discrepancies or differences in the charted data were discussed and resolved through consensus. In cases where consensus could not be reached, additional input from the review team was sought to make the final decision. This process ensured that the final dataset was reliable and agreed on by multiple reviewers. After charting data from each included study, the reviewers entered the extracted information into a predesigned data extraction form.

Results

The PRISMA flow chart diagram, as illustrated in figure 1, was used to systematically document the identification and selection of studies for this review,^{13 14} ensuring a rigorous and comprehensive selection process. Initially, 868 records were identified from PubMed (n=288) and Scopus (n=580). After removing 232 duplicate records, 636 records remained for title and abstract screening, leading to the exclusion of 403 studies based on relevance. We sought to retrieve 233 full-text reports, but 22 were unavailable. The remaining 211 reports were assessed for eligibility based on our predefined inclusion and exclusion criteria, resulting in the exclusion of 76 reports due to different interventions, 23 due to different outcomes and 8 because they were reviews. Ultimately, 104 studies met the criteria and were included in the review. A comprehensive table summarising all extracted information from the included studies is provided in online supplemental file 2. This table reports the study titles, countries where the research was conducted and their respective settings (field and/or laboratory), brief summaries of each study’s aim(s), the study design (observational, experimental or computational), the methodological approach, mosquito population characteristics and total sample size, the description and area(s) of intervention or technologies employed or developed, as well as the reported outcomes and key results.

Figure 1. PRISMA chart diagram of the study selection process. The chart details the number of records identified from databases, the number of records screened, the reports assessed for eligibility and the final number of studies included in the scoping review.

Study type based distribution of papers

The distribution of the papers included in this review according to the type of study is as follows: 82 (79%) evaluative studies, 15 (14%) experimental studies, 3 (3%) cross-sectional studies, 3 (3%) observational studies and 1 (1%) simulation study.

The predominance of evaluative studies can be explained by the fact that most of the technologies used were developed and tested by the authors themselves. Newly developed technological methods were then assessed by comparison with conventional manual methods, as reference. However, experimental, cross-sectional and observational studies are some applications and implementations of these previously developed methods in order to: (1) examine the influence of specific factors on the accuracy and efficiency of the developed technological tools (experimental studies); (2) real-time characterise and map specific localities for example (cross-sectional studies); and (3) observe phenomena such as the spatiotemporal distribution patterns and feeding behaviour of mosquitoes (observational studies).

Geographic distribution of the studies

The reviewed studies reveal a geographical distribution across multiple regions, as shown in figure 2. Predominantly, research is concentrated in the USA, with numerous studies originating from various states. Significant contributions also come from Brazil, where extensive research efforts focus on mosquito-borne diseases, as well as Malaysia, which shows a strong commitment to vector management strategies. Countries in Africa, including Burkina Faso, Tanzania and Zimbabwe, account for a notable number of studies, reflecting active research efforts in this region. In Asia, notable research efforts emerge from India alongside contributions from Singapore and China. European nations such as the UK, Spain and France also feature prominently, reflecting a global interest in innovative mosquito management strategies. Additionally, some studies encompass multi-country collaborations, highlighting the interconnected nature of vector control research.

Figure 2. Geographical distribution of the included studies. The area of each circle is proportional to the number of studies conducted in that country. Labels are shown for the top 10 countries with the highest number of studies.

Targeted mosquito population

The studies highlight a significant focus on various mosquito species crucial for vector-borne disease transmission. Dominant species include Aedes aegypti and Aedes albopictus, key vectors for diseases such as dengue and Zika, alongside Anopheles gambiae and Anopheles arabiensis, which are vital in malaria transmission. Additionally, studies frequently address Culex sp, such as Culex pipiens and Culex quinquefasciatus, associated with West Nile virus. The presence of mixed-species populations highlights the ecological diversity in the regions studied and underscores the need for targeted vector control strategies that consider the unique behaviours and habitats of these species.

Overview of study designs and technological advances in mosquito control and monitoring

The study’s findings demonstrate a wide array of technological innovations being used in mosquito control and monitoring. Remote sensing and imaging include the use of drones (eg, DJI Phantom 4 Pro) and satellite imagery (eg, Landsat-8 TIRS) for capturing high-resolution data on mosquito habitats. Monitoring and trapping technologies encompass smart traps and ovitraps designed for effective population monitoring. In the domain of data analysis and modelling, statistical methods and machine learning models, such as random forest (RF) and support vector machines (SVMs), are used to predict mosquito populations and outbreaks, often in conjunction with GIS tools for spatial analysis. Finally, advanced techniques, including deep learning models and automated vision systems, enhance species identification and monitoring capabilities. The detailed technologies used under each category are discussed in the respective sections. While a few technologies dominate the research landscape, there is significant interest in exploring novel approaches, particularly in machine learning and deep learning. A significant majority of these studies employed experimental designs, often in conjunction with computational methods, highlighting a dual approach that combines empirical and algorithmic techniques. This methodological framework, which included field data collection alongside advanced computational analyses such as machine learning algorithms, effectively addressed complex ecological questions related to mosquito species and their behaviours.

Distribution of studies by research focus area

The distribution of studies by research focus area, as illustrated in figure 3, highlights the central role of AI and sensor-based technologies in smart mosquito surveillance systems. Pure AI applications represent the largest category, with 35 studies (33.65%), reflecting the growing reliance on AI for automated mosquito detection, classification and predictive modelling. AI for environmental monitoring follows as the second most prominent category with 24 studies (23.08%), while AI combined with sensors accounts for 19 studies (18.27%), demonstrating active efforts to integrate real-time sensor inputs with AI algorithms to improve detection accuracy and operational efficiency. Sensor-based approaches for environmental applications comprise 19 studies (18,27%), underscoring the continued interest in understanding mosquito dynamics through environmental data. In contrast, pure sensor-based systems constitute only seven studies (6.73%), indicating a more limited focus on traditional sensor-only approaches compared with AI-driven or hybrid methodologies. Overall, these results confirm the predominance of AI-centred and AI-assisted technologies in the development of next-generation mosquito surveillance systems.

Figure 3. Distribution of studies by research focus area. Distribution of studies by research focus area. The bar chart shows the total number of included studies (x-axis) categorised by the five primary research focus areas (y-axis). AI, artificial intelligence.

The discussion in the following subsections is organised according to the main research focus areas identified. Each subsection provides a detailed analysis of the studies belonging to that specific category. References are cited collectively within the relevant discussion, meaning that the number of references reported corresponds to the total set of articles contributing to that research area, while individual authors are mentioned only when needed for explanatory purposes.

AI applications

The research covered in this category spans multiple countries, focusing on regions where mosquito-borne diseases are prevalent, often in Asia, Southeast Asia and other tropical and subtropical areas. For instance, many studies were conducted in the Philippines,15,17 South Korea^{18 19} and India.20,24

Although the dataset focuses primarily on these countries, other studies are set in a range of geographical locations, suggesting a global application of AI models developed for mosquito surveillance. However, instead of collecting or breeding mosquitoes, some works²⁵ used publicly available datasets collected by other researchers in previous studies.

The reviewed studies focus on a range of mosquito genera and species, each linked to specific vector-borne diseases. The genus Aedes is widely represented across the studies26,31 due to its relevance in transmitting diseases such as dengue, Zika and chikungunya. The two most frequently studied species within this genus are A. aegypti, commonly referred to as the yellow fever mosquito and a major vector for dengue and Zika viruses and A. albopictus, also known as the Asian tiger mosquito, another significant vector for dengue and chikungunya. The genus Anopheles, the primary vector for malaria, is also well represented in the studies.^{20 32 33} Key species include A. gambiae, known as the African malaria mosquito, and Anopheles stephensi, often found in urban environments. The genus Culex, responsible for transmitting diseases such as West Nile virus and lymphatic filariasis, is featured in some studies.³⁴ However, most papers involve more than one genus, such as Anopheles and Aedes,¹⁸ Aedes and Culex,2124 35,37 or all three together.15,1722 23 25 38 This is particularly true for papers focused on discriminating between different genera or species.

In fact, the majority of the retrieved works in this category involved the development of AI systems to discriminate among mosquito genera and species.15,1719 21 23 Most of these works focused on multiclass classification problems, while only a few addressed binary tasks. For instance, Pataki et al²⁷ focused on identifying tiger mosquitoes (A. albopictus) from pictures taken by citizens, while Asgari et al⁴⁰ explored their developed model in discriminating between A. aegypti and C. quinquefasciatus and between males and females. Sex identification was another common task,^{32 36 37 43 45} as distinguishing females is essential for targeting efforts to control the spread of mosquito-borne illnesses. Other classification tasks studied in the last decade included the detection of larvae in water,^{26 47} strain and storage methods,⁴³ and age reported as the gonotrophic cycle stage in female mosquitoes.²² All these works generally report results using classic classification metrics, that is, accuracy, recall, precision and F₁-score. The reported accuracies for species classification using deep learning models varied depending on dataset size and image quality, generally ranging from 76.2%³⁷ to 99.83%³⁰ across the reviewed studies. Other tasks falling into the regression category, such as predicting mosquito abundance,¹⁸ trap counts³⁴ and counting the number of eggs,³¹ reported results in terms of errors such as mean absolute error and root mean squared error. A distinct regression approach is presented by Tan et al,⁴⁸ who use a deep neural network to predict Aedes egg counts and future dengue outbreaks, specifically integrating not only dengue cases but also historical entomological data and meteorological data. In most of these studies, researchers employed images of mosquitoes either on plain backgrounds or in more complex and realistic environments. Some extracted specific features, such as wing geometric morphometry,⁴² while others used smartphone camera captures.^{20 22} Additionally, some exploited acoustic features of mosquito wingbeats,^{15 16} converting the information into spectrograms,^{25 36 46} that is, into images. Consequently, most of them leveraged supervised learning of deep learning techniques, mainly computer vision methods. In particular, convolutional neural networks (CNNs) were prominently applied.^{36 40 47} Most works employed well-known architectures such as AlexNet,^{37 39} ResNet^{25 27 30 31 46 49} and VGG,1923 24 40,42 49 and the transfer learning paradigm, in which models trained on other data are fine-tuned and adapted to the new task and data at hand.^{19 23 26 27 35 40 43 46} Others have employed either more recent architectures based on transformers⁴⁵ or more traditional machine learning techniques. For instance, Methiyothin and Ahn²⁸ employed XGBoost, and Chinnathambi et al³⁴ used a hybrid solution with the support vector machine-partial least squares regression model, while others^{15 16 29} have employed statistical methodologies.

Furthermore, apart from performance metrics, some studies proposed and analysed the impact of preprocessing techniques,¹⁵ including methods for reducing class imbalance²⁴ and distilling knowledge into more compact models.¹⁷ They also explored the explainability of the developed models,^{19 32 40} a topic that has gained significant attention in recent years in AI research. Overall, the reviewed studies highlight the significant potential of AI in advancing mosquito surveillance and control efforts globally, facilitating the work of researchers in discriminating between mosquito genera and species, predicting population dynamics, and identifying key characteristics such as sex and age.

AI for environmental applications

The research in this dataset encompasses a global response to mosquito-borne diseases, with studies conducted in various regions such as the USA,50,52 Brazil,^{53 54} Thailand,^{55 56} Kenya,⁵⁷ Burkina Faso,^{49 58} China,^{59 60} India,⁶¹ Greece,⁶² Hungary, Slovenia and Serbia,⁶³ Peru,⁶⁴ South Korea,⁶⁵ Spain,⁶⁶ Sri Lanka,^{67 68} Zimbabwe⁶⁹ and Malawi.⁷⁰ These studies demonstrate the widespread application of intelligent environmental monitoring systems and smart mosquito surveillance technologies for effective mosquito tracking and disease control.

A significant focus of the research is on the Aedes genus, particularly A. aegypti and A. albopictus, primary vectors for dengue, Zika and chikungunya.5153,55 59 Other studies investigate the Anopheles genus, responsible for malaria,4957 58 69,71 and the Culex genus, linked to West Nile virus and lymphatic filariasis.^{65 72} The studies showcase various AI and environmental monitoring approaches to tackle mosquito-borne diseases employing a diverse range of technologies, categorised into different key areas. Several studies leverage Remote sensing technologies including drones such as the DJI Phantom 4 Pro and 3DR Solo, which capture high-resolution RGB and multispectral imagery essential for monitoring mosquito breeding sites.^{49 54 57 64 67 68 70 71} Furthermore, image processing and computer vision techniques, such as CNN and YOLO V.7, were used for effective habitat identification and classification.⁶⁰ Following the same principle of the importance of detecting mosquito breeding sites, Bhutad and Patil⁶¹ developed a cloud-based deep learning system in India that identifies mosquito breeding hotspots using YOLO V.3, based on the detection of stagnant water via public camera infrastructure.

The application of machine learning algorithms such as RF and XGBoost allows for accurate predictions of habitat suitability and mosquito population dynamics,^{63 72} while SVMs were applied to classify climate data related to disease outbreaks.^{53 55}

Additionally, Stanton et al have integrated GIS to facilitate spatial analysis to identify malaria hotspots and map mosquito habitats through Geographic Object-Based Image Analysis.⁷⁰ A similar approach was used by Porcasi et al⁷³ in Argentina, who applied statistical spatial analysis to identify ‘high-activity clusters’ of A. aegypti. Their study analysed egg counts from ovitraps and correlated these clusters with urban cover types derived from Sentinel-2 satellite data, providing a tool for optimising surveillance strategies. These integrated approaches highlight the potential of combining advanced technologies to enhance surveillance and control strategies for mosquito-borne diseases.

The reviewed studies yielded significant outcomes regarding the classification and prediction capabilities related to mosquito surveillance and control. High classification accuracy was achieved for identifying water bodies and mosquito habitats, with advanced techniques such as Mask R-CNN demonstrating a precision of 85.2% in urban areas, despite some challenges with overlapping objects that reduced overall accuracy to 77.5%.⁵⁶ Quantitatively, the predictive models for vector abundance and disease outbreaks demonstrated high reliability, with area under the curve (AUC) values generally ranging from 0.808⁶⁵ to 0.99.⁵⁵ Similarly, automated identification of mosquito habitats using drone imagery and satellite data achieved classification accuracies typically between 86.73%⁶⁴ and 98%.⁷¹ The studies established clear correlations between the number of adult mosquitoes and larvae with environmental factors such as temperature and rainfall.^{53 62 73} Predictive models, particularly using RF algorithms, effectively forecasted annual mosquito infection rates and the spatiotemporal dynamics of A. aegypti populations, emphasising the critical role of prevention practices in outbreak preparedness.^{51 55} Additionally, habitat suitability predictions for A. albopictus based on climate scenarios were made, while models incorporating environmental factors such as NDVI (Normalized Difference Vegetation Index), land surface temperature and water area enhanced understanding of mosquito abundance.⁶² Overall, these findings highlight the effectiveness of integrating advanced technologies and environmental data to inform targeted control efforts for mosquito-borne diseases, in addition to demonstrating the potential of AI and environmental monitoring technologies in combating mosquito-borne diseases. By exploring diverse geographical locations and mosquito species, these researches contribute to the development of effective surveillance and control measures.

Sensor-based technologies for environmental applications

The studies included in this section emphasised diverse sensor technologies and remote sensing methods applied to environmental monitoring, with a focus on mosquito populations and vector control, and with a notable concentration of studies originating from Brazil, Malaysia and the USA. A recurring theme across several papers is the use of drone technology. For example, Valdez-Delgado et al⁷⁴ and Byrne et al⁷⁵ employed drones in Mexico and Côte d’Ivoire, respectively, to identify potential breeding sites for A. aegypti and A. funestus. Similarly, Hardy et al⁷⁶ used drones in Zanzibar to map malaria vector habitats, highlighting the effectiveness of aerial imaging for larval source management. These studies underscore the precision and cost-effectiveness of drone-based surveys, particularly in hard-to-reach areas.

Study by Aldosery et al⁷⁷ developed IoT-based systems for mosquito monitoring by implementing Mosquito Ovitraps IoT Sensing System in Portugal, using a combination of weather and water sensors with a real-time connection to the cloud for data transmission which offers real-time insights into mosquito breeding sites, significantly improving the timeliness and accuracy of data collection for vector control. Moreover, monitoring environmental factors such as temperature, humidity and rainfall is necessary, as they were highlighted as key determinants of mosquito populations by several studies. For instance, Benedum et al⁷⁸ developed the PLUM (Predictive fLUshing-Mosquito) model in Singapore to predict dengue outbreaks based on rainfall patterns. In Saudi Arabia, Jemal and Al-Thukair⁷⁹ mapped mosquito breeding sites and correlated mosquito abundance with climatic factors, while Heath et al⁸⁰ in French Polynesia modelled the impact of environmental heterogeneity on mosquito dispersal, based on data collected via traps. Caputo et al⁸¹ in Italy combined citizen science-facilitated field data collection with geostatistical analysis of remote sensing data. Their study used multispectral aerial imagery, LANDSAT-8 satellite data to estimate surface temperature and LiDAR (Light Detection and Ranging) data to create continuous predictive maps of A. albopictus egg distribution. Similarly, Lorenz et al⁸² in Brazil used remote sensing techniques, including thermal images and landscape features, to predict A. aegypti infestation hotspots. These studies emphasise the need for integrating environmental data into vector control strategies.

Several studies were included that fit the ‘Technologies’ search criteria by applying advanced prediction models to environmental factors, often sourced from remote sensing. For example, Tonnang et al⁸³ predicted shifts in malaria vector distribution under future climate scenarios using the combining of CLIMEX platoform and ArcGIS systems, while Little et al⁸⁴ used climate-based models to predict West Nile virus infection rates in Culex mosquitoes in the USA, obtaining meteorological data with the North American Land Data Assimilation Systems. Similarly, Espinosa et al⁸⁵ in Argentina combined landscape analysis, performed on images obtained from SPOT 5 satellite and spatial pattern modelling to identify dengue breeding sites. In Austria, Zittra et al⁸⁶ examined CORINE Land Cover images to determine the effects of landscape structure and climatic factors on mosquito communities, finding that land cover and seasonality were significant determinants of mosquito species distribution and abundance.

Some studies focused on low-cost, innovative technologies. Pan et al⁸⁷ used a Raspberry Pi microcluster for environmental data collection in mosquito research applications. Other approaches included the development of low-cost optical imaging systems for remote surveillance⁸⁸ and infrared sensors for bioacoustic species identification.⁸⁹ While Cheong et al⁹⁰ developed the PesTrapp mobile app in Malaysia, which facilitates real-time field and laboratory monitoring of mosquitoes. The app uses GPS to record the coordinates of trap locations, providing precise data for mosquito monitoring and control efforts.

Chaves et al conducted research in Costa Rica to assess the association between A. aegypti ovitrap egg counts, multiscale remotely sensed environmental data and arboviral cases.⁹¹ The study found that mosquito egg counts increased with temperature SD and decreased with Enhanced Vegetation Index. The use of ‘ovitraps’ allowed for the identification of high-risk areas for disease transmission, providing a low-cost and easily deployed technique for mosquito surveillance and control.

Lastly, the development of protocols for integrating multiple methods is evident. For example, Byrne et al⁷⁵ created a workflow that integrates drone surveys and entomological sampling to analyse Anopheles breeding habitats in agricultural landscapes. Similarly, Vacek et al⁹² designed small, unmanned aircraft systems for environmental sensor deployment and retrieval, offering a versatile tool for pest monitoring and data collection. In conclusion, these studies demonstrate a broad range of sensor technologies, remote sensing methods and computational models that are applied to mosquito monitoring and control. From drones and IoT systems to machine learning and low-cost devices, the research highlights innovative ways to enhance vector surveillance, environmental data collection and public health efforts in various ecological and climatic settings.

AI and sensors

The conducted search highlighted a substantial body of research focused on AI and sensor technologies for smart mosquito surveillance, with a notable concentration of studies originating from the USA,93,97 Brazil,^{98 99} and Taiwan.^{9 100 101} The geographical distribution covers a wide range of regions, including Africa,¹⁰² Europe^{103 104} and Asia,102105,108 where mosquito-borne diseases are a significant public health concern. Overall, this extensive body of work reflects a global commitment to advancing technology in public health initiatives aimed at controlling mosquito populations and mitigating disease transmission.

Aedes and Culex mosquito population are predominantly targeted,991,93 100 with A. aegypti and A. albopictus highlighted as primary vectors for diseases such as dengue and Zika. Culex sp, particularly C. quinquefasciatus and C. pipiens, are also prominent due to their involvement in transmitting West Nile virus and other pathogens. Some studies exclusively focus on Aedes mosquitoes^{96 99 101 106} given their critical role in urban disease transmission. Additionally, Anopheles sp, including A. arabiensis and A. gambiae, are highlighted for being primary vectors of malaria.^{97 102 105}

The reviewed studies demonstrate the use of advanced sensor technologies and AI in mosquito surveillance and smart trapping systems. Key technologies include Finite State Machines for modelling mosquito flight dynamics by Li and Principe,⁹³ 3D printing for creating components such as the Smart Mosquito Zapper by Huang et al¹⁰⁰ and smartphones, such as the Samsung Galaxy S5, for capturing mosquito images.^{94 105} Additional sensor technologies include infrared and laser sensors for detecting wingbeat patterns and flight data,^{98 101} as well as optical sensors with light sources and photo transistors for flying insect detection.^{99 103} AI-based solutions feature deep learning models such as AlexNet and fully convolutional networks for mosquito detection,¹⁰⁷ alongside TensorFlow Inception V.3 for species identification from mosquito wingbeat recordings by Khalighifar et al.⁹⁷ unmanned aerial vehicles (UAVs) equipped with CNNs are used to classify mosquito habitats.⁹⁹ Several smart trap systems also integrate real-time image recognition and processing, using platforms such as Raspberry Pi and SqueezeNet neural networks,⁹ and devices such as MosID for high-accuracy species identification through computer vision.¹⁰⁹ Study conducted by Semwal et al relying on GPU (Graphics Processing Unit)-enabled robots such as the Dragonfly,¹⁰⁸ employing UV (ultraviolet)-powered traps and differential drive mechanisms for real-time mosquito classification. These technologies collectively enhance the accuracy and efficiency of mosquito surveillance and control efforts whereas the study by Cavagna et al¹¹⁰ used 3D-video tracking approaches and computer vision algorithms in Italy to study mosquito swarming behaviour.

Mosquito species identification, classification, and accuracy in recognition are the major outcomes of the studies reviewed under this category. The primary metric used across these studies is accuracy, generally expressed as a percentage, to evaluate the correctness in identifying mosquito species, genders or other characteristics. Many studies focus on differentiating between species, such as A. aegypti and Culex mosquitoes, as well as distinguishing between local and invasive species.^{9 93 94 98 100 101 104 105 108} Specific tasks include identifying gravid female mosquitoes and recognising wingbeat characteristics.⁹⁵ Across the reviewed literature, integrated AI-sensor systems consistently demonstrated high diagnostic precision, with reported accuracies for species and sex classification generally ranging from 80%^{94 96} to over 98%¹⁰⁰ depending on the complexity of the dataset. In addition to accuracy, other metrics such as AUC, precision, and recall are employed to assess the effectiveness of various identification methods,^{96 99 109} including optical sensors and classifiers. For example, studies examining mosquito egg identification mention error rates compared with traditional methods, and some evaluations also extend to performance metrics such as the F1 score and the effectiveness of counting mosquitoes. Overall, these studies explore a range of tools and approaches, from UAVs combined with neural networks to smartphone-based species detection systems, aiming to optimise and validate mosquito recognition and classification techniques across different environments and species.

Sensor technologies

Studies under this section have used sensor technologies and chemical interventions for mosquito detection, monitoring and control, focusing on methods to address vector-borne disease transmission.

Che Dom et al developed an IoT-based Ovitrap system using NodeMCU for the detection of Aedes mosquitoes, the primary carriers of Dengue fever.¹¹¹ This research, conducted in Malaysia where dengue fever is widespread, involved modifying the existing sticky ovitrap to integrate IoT components. Data from the ovitrap were transmitted directly to the Blynk mobile application, enabling remote monitoring. The system proved effective, and after adjusting the entrance size of the trap and the battery supply, a functional model capable of detecting mosquitoes in real-time was created. The ‘ovitraps’ allowed identification of areas where disease rates were higher and acted as a low-cost and easily deployed technique. Similarly, Zulkiply et al¹¹² introduced a smart mosquito trap capable of sensing the nearby mosquitoes and transmitting real-time data on the average number of trapped specimens to the Favoriot IoT cloud platform for remote monitoring. The system uses a Carbon dioxide attractant made from a mixture of yeast and brown sugar to lure mosquitoes into the trap.

Hamdan et al conducted a study in Malaysia, using ovitraps to monitor Aedes mosquito populations.¹¹³ The study used ovitraps equipped with sensors to continuously monitor mosquito populations, providing detailed insights into the spatial distribution and breeding preferences of Aedes mosquitoes.

Measuring and monitoring wingbeat frequency and swarming behaviour is another key element when it comes to vector control. In this regard, Staunton et al¹¹⁴ in Australia employed ultrasonic Doppler sensors to track mosquito wingbeat frequencies. Similarly, De Nadai et al developed optical sensors with infrared LEDs (Light-Emitting Diodes) to measure the wingbeat frequencies of mosquitoes, allowing for species identification based on these frequencies. This approach is particularly useful for automated species identification in surveillance systems, enhancing the accuracy and efficiency of mosquito monitoring.¹¹⁵ Building on this concept, Parker et al¹¹⁶ used infrared video tracking, but with the aim to study mosquito interaction with ITNs (Insecticide-Treated Nets) in the UK, emphasising the impact of bed net design on mosquito behaviour. On the other hand, Voloshin et al¹¹⁷ in the UK used retroreflective imaging to improve video tracking of mosquito flight paths.

These studies collectively highlight the potential of integrating sensor technologies to enhance mosquito detection, monitoring and control, offering scalable and cost-effective solutions to combat vector-borne diseases globally.

Economic and social acceptability considerations

The successful transition of advanced mosquito surveillance systems from experimental prototypes to scalable public health tools relies not only on entomological precision but fundamentally on economic viability, operational feasibility and social acceptance. Our review reveals that while technological capabilities have advanced rapidly, the evidence regarding the cost-effectiveness and practical deployment of these systems remains fragmented. Economically, as detailed in online supplemental table 1, many included studies qualitatively characterise their solutions as ‘low-cost’ or ‘affordable’, particularly those leveraging off-the-shelf consumer electronics such as Raspberry Pi microclusters⁸⁷ or smartphone-based acoustic sensors.¹⁰² However, formal cost-utility analyses comparing these high-tech interventions against traditional manual surveillance are largely absent. Despite this, operational efficiencies were frequently highlighted as a proxy for cost savings; for instance, the integration of drone imagery with targeted entomological sampling has been shown to significantly reduce field effort and labour hours,^{74 75} while mobile applications such as PesTrapp streamline data collection workflows, implying reduced long-term operational expenditures.⁹⁰

Beyond direct financial implications, the ‘hidden’ costs of deployment, specifically technical complexity, training requirements and maintenance, emerge as critical determinants of success. High-tech solutions often introduce a trade-off between data granularity and the need for specialised human resources; for example, drone-based surveillance in regions such as Malawi necessitates trained local pilots and data analysts, creating a barrier to entry in resource-limited settings.⁷⁰ Conversely, citizen science initiatives attempt to mitigate these logistical burdens by designing user-friendly interfaces that require minimal training, thereby shifting the workload from professional entomologists to the community.⁴⁴

However, the sustainability of such community-based or residential surveillance systems is inextricably linked to social acceptance, as summarised in online supplemental table 2. The deployment of hardware in private spaces requires a high degree of public trust and engagement. Studies employing ovitraps or acoustic sensors in residential areas¹⁰² highlight that without ongoing community education, such as the ‘Communication for Behavioural Impact’ approach adopted in Malaysia¹¹³, surveillance efforts are prone to failure due to lack of participation or device tampering. Notably, Caputo et al⁸¹ demonstrated that proactive synergies with local administration and transparent communication can result in acceptance rates exceeding 93%, proving that ‘social investment’ is as crucial as technological investment. Therefore, to translate these innovations into policy, future research must prioritise holistic feasibility studies that rigorously evaluate maintenance costs, training scalability and community integration alongside entomological efficacy.

Discussion

This scoping review highlights significant advancements in technologies for mosquito surveillance and control. The adoption of smart surveillance systems including AI, GIS and IoT marks a shift from traditional, labor-intensive mosquito monitoring methods to automated data-driven approaches. These technologies are increasingly deployed across diverse geographies, from tropical regions with high endemic mosquito populations to temperate zones experiencing outbreaks due to climate and environmental changes. By synthesising the current body of literature, this review reveals how these technologies enhance species identification, support real-time monitoring, and improve habitat and outbreak predictions, addressing major public health needs in regions burdened by malaria, dengue and Zika virus.

One of the most significant advancements is the integration of AI for species identification, classification and behavioural analysis. Deep learning models such as CNNs can now process large image datasets to distinguish between mosquito species including A. aegypti, Culex and Anopheles. This innovation simplifies species identification and enables targeted vector control interventions. Traditional methods, which rely on manual sorting and microscopic analysis, are labor-intensive and require specialised expertise. In contrast, AI-driven systems offer faster, more accurate results, making them particularly valuable in resource-constrained settings where timely data is critical for disease control efforts. Crucially, the decisive advantage of AI models over traditional morphological methods lies in their operational scalability and rapidity. While manual identification is constrained by human throughput, AI algorithms can process vast datasets in real-time, enabling a level of temporal surveillance coverage that would be logistically unfeasible with conventional techniques.

Another key innovation is the integration of IoT-enabled traps and sensors. Smart traps equipped with infrared sensors and wireless data transmission allow for continuous, real-time monitoring of mosquito populations. These devices are particularly used in remote or densely populated areas where infrastructure for traditional surveillance is limited or nonexistent. IoT systems also enable remote data collection, allowing researchers to monitor activity and environmental conditions without physical site visits. When integrated with cloud computing, these systems can process and analyse large-scale datasets, offering insight into mosquito behaviour, breeding patterns and population density in near real time.

Environmental monitoring has been significantly enhanced through GIS and remote-sensing technologies, enabling the prediction and tracking of mosquito habitats based on climatic variables such as temperature, rainfall, and humidity. High-resolution imagery from drones and satellites is processed through predictive models to forecast mosquito population trends and assess the likelihood of disease outbreaks. This proactive approach is particularly crucial in regions undergoing environmental shifts, where climate change is expanding mosquito habitats and increasing the risk of vector-borne disease transmission in previously unaffected areas.

This scoping review benefits from an extensive search strategy and inclusive approach that spans various technologies and geographical regions. By incorporating studies on both established and emerging surveillance systems, it provides a comprehensive perspective on current innovations and trends in mosquito monitoring. However, the methodological limitations of this review must be acknowledged. First, the review’s scope has clear boundaries: the search was limited to English-language publications and conducted exclusively on the PubMed and Scopus search engines, which may underrepresent studies from certain regions or those indexed in other scientific databases. Second, the search strategy was deliberately focused on integrated digital systems; consequently, this review does not provide a comprehensive overview of all technological advances in mosquito surveillance, such as standalone chemical attractants or emerging molecular approaches such as eDNA. Finally, the inclusion criteria required studies to report at least one surveillance-related or control-related outcome, in order to focus on technologies that had been applied or tested within a mosquito surveillance context. While this choice ensured analytical coherence and relevance for public health applications, it may have led to the exclusion of very early-stage or purely conceptual technologies that have not yet been formally evaluated but could represent promising future directions.

These limitations suggest that further reviews focusing on empirical studies evaluating the effectiveness of these surveillance technologies in field settings are needed. Beyond these constraints, our analysis identifies critical gaps in the current literature. A fundamental gap is the discrepancy between technological development and real-world validation. A significant portion of the included studies primarily describes the development of prototypes or technological innovations, with limited documentation of field performance or efficacy. This highlights how many of these new technologies remain experimental and still require rigorous validation and better practical implementation.

This lack of validation extends to long-term sustainability and economic feasibility. We identified a notable absence of explicit focus on sustainability. Although our results include systems based on low-cost hardware (such as Raspberry Pi, smartphones and NodeMCU microcontrollers), future research must prioritise advances in miniaturisation and energy efficiency. The development of solutions such as solar-powered devices, which did not feature prominently in our review, will be essential for enabling long-term deployments in remote or resource-limited regions. Likewise, the long-term impact of these technologies on disease transmission, their cost-effectiveness and their adaptability in diverse environments remain unevaluated aspects, yet essential for optimising surveillance.

Furthermore, the transition from pilot studies to public health policy requires addressing the socioeconomic dimension of surveillance. While the proliferation of low-cost consumer electronics has lowered the barrier to entry for hardware, our review highlights a critical distinction between ‘device cost’ and ‘implementation cost’. The latter often includes overlooked expenses such as specialised personnel training and long-term maintenance, which can be prohibitive in resource-limited settings. In this context, citizen science models emerge not only as a strategy to reduce operational labour costs but as a vital mechanism for social integration. As evidenced by the high acceptance rates in studies prioritising community engagement, the sustainability of surveillance systems is contingent on public trust. Consequently, future implementation frameworks must treat social acceptance and community education not merely as prerequisites, but as integral components of the surveillance infrastructure itself, as critical as the sensors themselves.

Finally, significant operational barriers still hinder widespread implementation. The lack of standardised protocols for data collection, image annotation, and model evaluation limits comparability and reproducibility across studies, making it difficult to benchmark performance or aggregate evidence. Interoperability across sensing platforms, databases and public health information systems also remains limited, restricting seamless data integration and coordinated vector control responses. Furthermore, the use of geospatial and health-linked datasets raises important ethical and privacy considerations, particularly in settings where data governance frameworks are weak or inconsistently enforced.

Overcoming these challenges will require harmonised technical standards, initiatives for open and Findable, Accessible, Interoperable, and Reusable data, and strong cross-sector partnerships. Future research should also focus on developing multipurpose tools, also with the integration of multiple technologies into unified systems, for cross-species identification, enhancing attractants via specific light and sound frequencies and integrating motion sensors to improve trap accuracy. Prioritising interoperable platforms that integrate GIS, IoT and AI-based trap data will support real-time data sharing among public health sectors, enabling a coordinated response to the risks of vector-borne diseases.

Conclusion

The advancements in smart mosquito surveillance technologies outlined in this review offer transformative potential in addressing the growing threat of mosquito-borne diseases worldwide. By integrating AI, IoT and GIS with traditional entomological practices, these innovations enable more precise, efficient and scalable vector control strategies. This review highlights not only technological progress but also its profound public health implications, particularly in regions heavily burdened by malaria, dengue and Zika. However, key challenges remain, including the need for cross-species identification tools, improved miniaturisation and energy efficiency, and interoperable platforms for data integration. Addressing these gaps through future research could further optimise surveillance systems. As these technologies continue to evolve, they hold great promise for enhancing mosquito control efforts and reducing the global burden of mosquito-borne diseases.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.