基于小波包非对称分解的钢丝绳图像边缘检测

Image edge Detection of Steel Wire Rope Based on Wavelet Packet Asymmetry Decomposition

针对煤矿视频监控系统中钢丝绳图像边缘检测的问题,提出一种边缘检测改进算法。首先将图像进行小波包分解,以获得对图像不同结构特征的描述;进而对经小波包分解所得到的高频部分加以分析,保留部分空间,其他空间继续分解;然后采用模极大值理论进行边缘检测。研究结果表明,此算法能够较好地检测出钢丝绳图像中的细节特征,且检测速度较快。

Based on problems of image edge detection of steel wire rope on coal mine monitoring system, an improved algorithm of numerical image edge detection is proposed. The first step is to decompose the image in wavelet packet to describe different structure features of the images; then the high part of the frequency is ana- lyzed with partial spaces reserved and other spaces continued to be decomposed. This is followed by edge detection of image based on the theory of maximum module. The result shows that this algorithm can detect the detailed features of the steel wire rope image more rapidly.