基于CNN的彩色图像边缘提取方法

Edge detection of color image based on CNN

根据人眼对R、G和B的敏感特性,提出一种基于视觉感知的矢量角度彩色空间距离度量方法,较传统方法补偿了RGB颜色空间的不均匀性,有效提高了RGB颜色空间内色差度量的精确程度.对基于细胞神经网络(Cellular Neural Network,CNN)进行的彩色图像边缘提取进行理论分析,确定与其功能相适应的最佳模板参数.实验结果表明:该算法不仅实现了彩色图像的边缘提取,同时提取效果更全面,更符合人眼的视觉效果,具有较强的鲁棒性.

According to the sensitivity of human eyes to R, G and B, a distance measurement method based on angle of vectors for color space is proposed based on visual perception. This method improves the ununiformity of RGB color space compared with the traditional methods, and effectively increases the precision of color difference measurement in RGB color space. Edge detection of color image based on cellular neural network （CNN） is analyzed theoretically, and the opti- mal template parameters fit for its function are determined. The experimental results indicate that this method accomplishes the edge detection of color image, obtains the more comprehensive effect of detection, is more in line with the perception of the human eyes, and has extensive robustness.