Algorithm 1 Detection and segmentation of invasive grasses using high-resolution RGB images.
	Required: orthorectified image set I. Representative samples set G. Sample masks set H
	Training
1:	for $i\leftarrow 1,n$ do	▹$n=$ total images in G (labelled data)
2:	Load $G_i$ and $H_i$ images
3:	Convert colour space of $G_i$ into HSV
4:	Insert each colour channel into a feature array D
5:	Use 2D filters on $G_i$ and insert their outputs into D
6:	From $G_i$ and $H_i$, filter only the pixels with assigned labelling on D
7:	end for
8:	Split D into training data $D_T$ and testing data $D_E$
9:	Create a XGBoost classifier X and fit it using $D_T$
10:	Use K-fold cross validation with $D_E$	▹ number of folds = 10
11:	Perform grid search to tune X parameters
	Prediction
12:	for $i\leftarrow 1,m$ do	▹$m=$ total images in I
13:	Load $I_i$ image
14:	Convert colour space of $I_i$ into HSV
15:	Scan every pixel and predict the object using X
16:	$O_i\leftarrow$ Convert the data into a 2D image
17:	Export $O_i$ into TIF format
18:	end for
19:	return $O_i$