论文部分内容阅读
针对不同的光照以及不同天气条件引起叶片图像颜色不一致的问题,采用10种典型的颜色恒常性算法对葡萄叶片图像中颜色的恢复效果进行对比试验,用色差ΔEa*b*和估计光源角度误差2种方法对不同算法的颜色恢复效果进行评价。试验结果表明:各种算法对图像中叶片颜色恢复均有一定的效果,其中2nd-order Grey-Edge和Edge-based Gamut Mapping 2种算法对不同光照和天气条件下以及多个品种的葡萄叶片图像处理后,其色差误差和估计光源角度误差均达到最小;从视觉角度上看,这2种算法处理后的图像也最接近基准图像。试验还表明:这2种颜色恒常性算法对于解决农业领域中许多由于成像条件的差异所引起的图像颜色偏差问题有明显的效果,而且能很好地实现葡萄叶片图像中颜色的恢复。
According to the different illumination and inconsistent color of leaf images caused by different weather conditions, 10 typical color constant algorithms were used to compare the color restoration effects of grape leaf images. Using the color difference ΔEa * b * and the estimated light source angle error 2 The method evaluates the color recovery effects of different algorithms. The experimental results show that various algorithms have some effects on the color recovery of the leaves in the images. The two algorithms, 2nd-order Gray-Edge and Edge-based Gamut Mapping, After processing, the error of color difference and the error of estimated light source angle are minimized. From the visual point of view, the images processed by these two algorithms are also the closest to the reference image. Experiments also show that these two color constant algorithms have obvious effect on solving the problem of image color deviation caused by the difference of imaging conditions in the agricultural field and can well realize the color recovery in grape leaf images.