基于高效边缘检测的彩色图像灰度化算法  

Color Image Decolorization Algorithm Based on Efficient Edge Detection

作　　者：刘亚 张团善[1] 王恩芝 LIU Ya;ZHANG Tuanshan;WANG Enzhi(School of Mechanical and Electrical Engineering,Xi'an Polytechnie University,Xi'an 710613,China)

机构地区：[1]西安工程大学机电工程学院,陕西西安710613

出　　处：《轻工机械》2022年第6期52-58,共7页Light Industry Machinery

摘　　要：为了使彩色图像灰度化后能更好地保留原始对比度和细节特征,课题组提出一种基于高效边缘检测的彩色图像灰度化算法。课题组通过对线性参数模型产生的一系列备选图像采用像素差值网络(pixel difference networks, PiDiNet)实现鲁棒性、准确性的边缘检测,边界点数最多的备选图像所对应的权重则被选为映射函数的最优解,由于引用了像素差值卷积PDC,对边缘信息敏感度增强。实验结果表明:此算法相比于其他算法能较好地保留原彩色图像的细节特征,使得输出图像轮廓清晰且自然,主客观评价均较优。In order to better preserve the original contrast and detail features of color images after decolorization, a color image decolorization algorithm based on efficient edge detection was proposed. Robustness and accuracy edge detection was achieved by using pixel difference networks(PiDiNet) through a series of candidate images generated by the linear parameter model. The weight corresponding to the candidate image with the highest number of boundary points was selected as the optimal solution of the mapping function. Sensitivity to edge information was enhanced due to the introduction of pixel difference convolution(PDC). Experimental results show that compared with other algorithms, the proposed algorithm can preserve the details of original color images bette, and make the outline of the output image clear and natural, and have better subjective and objective evaluation.

关 键 词：边缘检测 灰度化 线性参数模型 像素差值网络 对比度保留 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN911.73[自动化与计算机技术—计算机科学与技术]