基于模拟退火算法的递归自动阈值分割方法

针对目标图像灰度对比度差的现象,以及对目标对象检测实时性的要求,并考虑到传统的Otsu分割方法在分割图像质量较差以及目标区域小时准确性差的缺点,提出了一种基于模拟退火算法的递归Otsu分割方法。在图像直方图呈双峰的情况下能够准确地找到分割阈值。在成像模糊、光照度较差的情况下此方法仍然可以获得较高的准确度。该方法在保证了检测质量的同时并没有导致运算时间的大幅度提升,有效地保证了处理的实时性。实际应用表明该方法切实可行。

Recursive Otsu Method Based on Simulated Annealing Algorithm

For the fast target detection in the poor gray - method based on simulated annealing algorithm is put forward. It is proposed to overcome the disadvantage of traditional Ot su method as low veracity when the pending image is poor and the target is small as well. A fine segmentation result could be expected even the pending image is fuzzy or obscure. With this algorithm, the image can be segmented effectively when lhe histogram distribute as a bimodal shape. With guaranteed detection results, no significant rising of time is caused, whic~x as- sures the real-timing of the process. The experimental results show that this method is practical and has better perlormance.