Algorithm 1. Obtaining Pixel Coordinates of Control Points in Test Images

Input: image block of static objects from test image

Procedure:

(1) Get reference image through identity of static object from database;

(2) Extract feature points for both test image block and reference image by SIFT operator [46];

(3) Perform feature matching to get homonymy feature point pairs;

(4) Employed RANSAC [47] to remove false matching points; the remaining matching points marked as $P_{test}$ for test image and $P_{ref}$ for reference image;

(5) Calculate homographic matrix $H_{homo}$ by solving formula below: $$P_{Test}=H_{homo}\times P_{ref}$$

(6) Estimate pixel coordinates of control points in test images CPT as following formula, $CP{T}_{ref}$ is the set of pixel coordinates of control points in reference images: $$CPT=H_{homo}\times CP{T}_{ref}$$

Output:$CPT$