局部背景边缘干扰下运动视频完整轮廓检测方法

针对传统相邻帧差算法在对轮廓检测过程中无法有效解决局部背景边缘干扰的问题,提出基于高速并行细化算法的运动视频完整轮廓检测方法.基于Sobel算子对运动视频轮廓进行粗检测,通过形态学后处理对粗检测获取的运动视频目标轮廓进行腐蚀运算,过滤其中的噪声,使图像边缘向内紧缩;采用两次膨胀处理恢复并增强运动视频目标图像的连通区域,引入高速并行细化算法对图像进行细化处理,采集图像骨架,并获取线条平滑的完整运动视频目标轮廓图像.结果表明,所提方法检测出的运动视频目标轮廓更为完整、清晰,并且检测效率高.

Complete contour detection method for moving video under local background edge interference

In order to solve the problem that the traditional adjacent frame difference algorithm cannot effectively solve the local background edge interference problem in the contour detection process, a complete contour detection method for moving video based on the high-speed parallel thinning algorithm was proposed. Based on the Sobel operator, the contour of moving video was roughly detected, and the corrosion operation for the contour of moving video object obtained with the rough detection was carried out with the morphological postprocessing, and the noise was filtered to make the image edge be tightened inward. With two times of expansion treatment, the connected region of image of moving video target was recovered and strengthened, and the high-speed parallel thinning algorithm was introduced to refine the image and acquire the image skeleton. In addition, the contour of complete moving video target with the smooth lines could be obtained. The results show that the contour of moving video target detected with the proposed method is more complete and clear, and the detection efficiency is high.