针对视频流图像的目标检测,基于帧间差分理论提出了一种实时多运动目标检测算法。根据视频图像在实时处理中的应用,基于FPGA硬件平台并行处理快与可编程的特性,在FPGA上实现算法,经过MATLAB环境算法验证,并在设计的FPGA视频图像处理系统进行实现。该系统主要由图像采集模块、图像缓存模块、图像处理模块及图像显示模块4部分组成,通过OV5640采集图像数据,再对数据进行差分、滤波、形态学等一系列算法处理,存入DDR3中,最后通过VGA进行显示。实验结果表明,该系统在一定测距范围内可稳定跟踪运动目标,可实时显示,多目标检测识别率较高,能实时检测视频流图像中的多运动目标。

Aiming at the target detection of video stream images,a real-time multi-motion target detection algorithm is proposed based on the theory of inter-frame difference.According to the application of video images in real-time processing,based on the characteristics of parallel processing of FPGA hardware platform,the algorithm is implemented in FPGA hardware,verified by MATLAB environmental algorithm,and implemented in the designed FPGA video image processing system.The system is mainly composed of four parts:image acquisition module,image cache module,image processing module and image display module.The image data is collected through OV5640,and then the data is processed by a series of algorithms such as difference,filtering and morphology,and is stored in DDR3,finally is displayed by VGA.The experiment results show that the system can stably track moving targets within a certain ranging range,can display in real time,has a high multi-target detection and recognition rate,and can detect multi-motion targets in video stream images in real time.