基于颜色恒常性的石材图像颜色校正的应用

Application of Stone Image Color Correction Based on Color Constancy

为了改善采集石材图像时因不同强度环境光照产生的图像颜色改变的问题,基于颜色恒常性原理,对石材图像进行颜色校正。运用欧氏距离计算Lab颜色模型下的颜色校正图像与标准图像的色差,使用MATLAB软件分析比较了Gray-World、 max-RGB、完美反射和图像熵约束等算法对于石材图像颜色校正的效果,并对完美反射算法和图像熵约束算法进行了改进。实验结果证实了改进的完美反射算法有效提高了石材图像的颜色校正效果,为石材图像接下来的图像识别、匹配和石材评价等过程提供了有效的基础。

In order to improve the problem of image color change caused by different intensities of ambient light when collecting stone images, the stone image is corrected based on the principle of color constancy. The Euclidean distance is used to calculate the color difference between the color corrected image and the standard image under the lab color model. The effects of Gray-World, max-RGB,perfect reflection and image entropy constraint algorithms on stone image color correction are analyzed and compared with MATLAB software, and the perfect reflection algorithm and image entropy constraint algorithm are improved. The experimental results show that the improved perfect reflection algorithm effectively improves the color correction effect of stone image, and provides an effective basis for the following processes of stone image, such as image recognition, matching and stone evaluation.