数字图象的最大熵直方图增强

常规的直方图均衡化图象增强方法会导致图象直方图的熵下降,从而损失了图象的某些信息,为此提出一种保熵增熵的增强方法。其主要思路是:首先提取并保存图象的细节信息,然后进行直方图谱线的兼并与推平处理,最后把细节信息再迭加上去。实验表明,这种方法,处理后的图象细节清晰,层次丰富,明显地优于常规直方图增强方法。

IMAGE ENHANCEMENT BASED ON MAXIMUM ENTROPY OF HISTOGRAM

The conventional enhancement with histogram would lead to lower the entropy of histogram; so that some image information would be lost. Presents a new enhancement method which could preserve or increase the histogram entropy and make the histogram appears to be nearly a even distribution in the whole grey field. The main thinking is: detecting and preserving the details of image; then annexing and leveling the histogram lines; eventually,superposing the details on the processed image. The experiment shows that the image enhanced by new method, in which the details is clear and the grey levels are rich,is superior than the image processed by conventional histogram enhancement.