Aiming at the problem of low efficiency of traditional painting robots

a smart painting system based on Kinect V2 image sensing equipment is designed.Firstly

the mismatched point pairs are eliminated by the RANSAC algorithm with adding Sampson operator to obtain accurate point cloud data of the painting area.Then

for irregular painting surfaces

the stereo bounding box method is proposed to find the nearest points of the two point sets

shortening searching time

therebythe point cloud slices are generated quickly.Finally

the normal vector of the data points are estimated to form the painting trajectory of the car scratch zone.The experimental results show that the system can complete the repair of the painting surfaces accurately and efficiently

as well as has a good application prospect.