讨论了灰度世界算法,提出了两种改进的灰度世界算法——基于标准差加权的算法和基于图像熵约束的算法,从而避免了当给定图像仅有少量几种颜色或大块单一颜色时的冥法失效问题.通过对算法颜色校正结果进行主观(视觉上)评价,并对算法性能通过使用色差数据进行了客观评价.实验结果表明：改进算法对实际采集的图像和下载的图像都有较好的校正效果.

The grey world algorithm is discussed in detail,and two improved grey world algorithms are proposed,to overcome the limit that the algorithm may be ineffective in the case of few colors or large union colors image which are based on standard deviation weighted algorithm and based on image entropy constraint algorithm respectively.The algorithms are evaluated in two aspects: the color correction results are compared subjectively (visually),and the performance is evaluated objectively using color difference data.The experimental results show that the improved algorithms have better correction effect on both real capture images and downloaded images.