基于灰度共生矩阵的自适应图像边缘检测

纹理分析是图像处理中一种十分重要的方法。通过纹理分析,利用灰度共生矩阵惯性矩特征值能够反映图像灰度空间复杂度的特性,成功获取了LOG边缘检测算子最佳空间系数,抑制了图像中的大部分噪声。并通过基于TMS320C6000专用信号处理器的图像处理系统实现了图像边缘检测的自动提取。实验结果表明,采用这种方法可以取得很好的噪声抑制效果,并且检测出的边缘精度高,实时性强,便于实现。

A Self-adaptable Method to Detect Edge in Images Based on Gray level co-occurrence matrix

Texture analysis is one of important methods in image processing. Characteristic number of inertia moment of gray level co-occurrence matrix can reflect the feature of complexity of image gray space. Optimum space coefficient of LOG operator can be ac-quired effectively and most noise of images can be removed. Edge detection in image is realized by image processing system based on TMS320C6000 DSP. The results show that this method eliminates noise efficiently, enhances the precision of detecting edges greatly and can be realized easily.