【摘要】 针对传统运动目标检测与跟踪算法计算量大、检测速度慢、鲁棒性差的问题，提出了一种通过动态调整检测区域以实现高效运动目标检测与跟踪的方法。利用改进的帧间差分算法提取运动目标，将目标大小和位置信息分别传给区域裁剪算法和卡尔曼滤波跟踪算法，实现检测范围的缩小和目标跟踪。采用改进后的方法对分辨率为640×480的图像进行实验，检测到目标的大小为66×84,调整后的检测区域大小为132×168,帧间差分检测部分计算量减小了约92.78%,并且根据目标位置进行跟踪的效果较好，在FPGA中验证了此方法的正确性。更多还原

【Abstract】 Aiming at the challenges of intricate computational requirements and a sluggish detection pace, and poor robustness of traditional algorithms for detecting and tracking moving objects, a proposed approach aims to attain efficient moving object detection and tracking by dynamically adjusting the detection area.Utilizing an enhanced inter-frame difference algorithm for the extraction of moving targets, the information regarding the size and location of the target are transmitted to the region cropping algorithm and the Kalman filter tracking algorithm, respectively, to achieve reduced detection range and target tracking.Using an improved method for a resolution of 640× 480 images are used for experiments, and a target size of 66 is detected to 64×84,the adjusted detection area size is 132 ×168,the computation of the inter frame differential detection part is reduced by 92.78%,and the tracking effect based on the target position is better, this approach has been confirmed to be correct in FPGA.更多还原