基于DSP/FPGA的嵌入式实时目标跟踪系统

Real-time Vision System for Tracking Object with DSP and FPGA Architecture

提出了一套基于DSP/FPGA的协处理器结构用以实现实时目标跟踪的嵌入式视觉系统。系统由DSP作为主处理器进行全局控制,利用具有流水线并行处理结构的FPGA作为协处理器实时完成DSP分配的处理任务。系统由FPGA快速完成最初的运动估计的结果,DSP在此基础上进一步分析和校正,并将校正信息反馈给FPGA,实现快速而准确的跟踪。

This paper presents an embedded monocular vision system for object tracking in real time with coprocessor architecture based on DSP and FPGA. DSP serves as primary processor to control the whole system flow, while FPGA as coprocessor with pipeline parallel processing. For rapidly and correctly tracking an object, the motion estimation of the current frame computed by FPGA is analyzed and verified by DSP, and then the verified result is fed back to FPGA for motion estimation in the following frame. The sonar sensor is used as complement to evaluate depth distance for obstacle avoidance.