提出了一种照明无关的小波多尺度相乘边缘检测方法,用于从非均匀的弱照明图像中提取边缘。根据照明反射图像形成模板与CCD相机成像原理,推导出图像的对应小波变换公式。然后,对图像局部区域中噪声、边缘与背景像素的小波系数进行比较分析,设计了一种照明无关的小波边缘检测公式。为增强边缘并抑制噪声,提出了一种改善的小波多尺度相乘边缘检测方法,并依照小波变换后边缘像素的特征,提取单像素的边缘。采用仿真和真实的非均匀的弱照明图像对该边缘检测算法进行验证,并与另外两种边缘检测方法进行定性的和定量的比较。实验结果证实了这种边缘检测方法能够从灰度不均匀的低衬比度图像中正确有效地提取边缘。

An illumination-independent wavelet scale multiplication edge detection method is proposed to detect edges from the non-uniform weak illumination image. Based on the illumination-reflected image formation model and CCD camera imaging formula, the image wavelet transform formula is deduced. The wavelet module coefficients of the background pixel, noise, edges are observed and analyzed in the local region, and the illumination-independent wavelet edge detection algorithm is worked out. To enhance edge pixels and reduce noise, an improved wavelet scale multiplication edge detection algorithm is applied to orient edges accurately. Through utilizing edge pixels' characteristics after wavelet transform, the single-pixel edges are taken out of the image. In the end, the edge detection experiments for synthetic and real non-uniform weak illumination images were conducted, and its performance is analyzed and compared with other two edge detection methods qualitatively and quantitatively. The experimental result proves it correct and effective when used in the non-uniform gray and low contrast images to detect edges.