Design and Implementation of Three Frame Difference Method and Camshift Algorithm Motion Object Detection System Based on FPGA

In response to the common occurrence of false edges and edge missing in traditional inter frame differential motion object detection algorithms, an improved three frame differential and improved Camshift algorithm motion object detection system was designed. The three frame differential and improved Camshift algorithm were integrated to detect motion objects, and a video image processing system was constructed on the Field Programmable Gate Array (FPGA) platform. The system mainly consists of five parts: image acquisition module, image caching module, three frame difference method, improved Camshift algorithm module, and image display module. Collect image data with a resolution of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$1024 \times 768$</tex> using an OV7670 camera, store the collected image data in SDRAM, and use an improved three frame differential algorithm to detect and label moving targets. Finally, track moving targets using an improved Camshift algorithm. The experimental results show that the system can accurately detect multiple moving targets in the video source collected by the camera and meet the requirements of real-time detection.