Elevator door protection method based on new moving image identification

Abstract

The elevator door protection system is one of the most important system in the elevator, the elevator door protection system based on image processing technology has good application prospects. Oriented to large storage capacity and strong dynamic characteristics of image technology in the process of elevator door protection, in this paper, a kind of elevator door protection system based on dynamic image recognition is brought forward. Modified median filtering method is established to filter the image noise and to hold on the boundary clearly. Otsu method is adopted for global threshold segmentation. Relying on background subtraction method and temporal difference method, the dynamic target identification model is set up. Combined with region growing segmentation algorithm, the dynamic target's movement is detected on time. The test platform is built up for the elevator door protection system. The experimental results show that the image identified process would be finished by detecting 3 frames of images. And the reliability of system is proved out with different velocity of moving.