摘要
针对铁路复杂环境的视频运动目标检测易受影响的情况,采用TMS320DM8168 DSP处理器,提出一种融合三帧差法和自适应的VIBE改进算法,采用SUSAN边缘检测算子、形态学算子处理、中值滤波后处理进行运动目标检测,并转化成HSV空间模型,将图像处理算法移植到DM8168平台上,在铁路沿线复杂的场景下,能够实时地准确检测出视频中运动目标,有效地消除空洞和鬼影,实现良好的鲁棒性。
For the video moving target detection in complex environment of railway is easily affected by the situation, adopts TMS320DM8168 DSP processor, presents a fusion of the three frame difference method and adaptive VIBE algorithm, uses SUSAN edge detection operator, morphological processing, median filtering and moving target detection, and transformed into a HSV space model, the image processing algo- rithm is transplanted to the DM8168 platform, along the railway complex scene, can accurately detect moving objects in video, effectively eliminate the empty and the phantom of the opera, to achieve good robustness.