3D Semantic VSLAM of Dynamic Environment Based on YOLACT

Algorithm 1

Dynamic information detection.

	Input: Get video streaming from the camera ;
	Output: Dynamic zone template Mask, Processed video stream ;
(1)	Use the Lucas-Kanade method to detect dynamic areas;
(2)	if then
(3)	(1) Intercept the same area as the previous frame to get ;
(4)	(2) Do ViBe+ dynamic detection of the current frame being intercepted with the previous frame
	;
(5)	else
(6)	;
(7)	end if