CAJ | 학술논문

针对传统局部阈值算法在分割低质量图像时存在分割时间长的问题，提出一种新的局部阈值分割算法。该算法首先去除传统阈值模型中局部标准差因式，再利用积分图像快速计算局部均值，然后利用求得的局部均值和像素点的灰度值来计算局部阈值。由于降低局部阈值计算的时间复杂度，从而减少了图像分割时间。实验结果表明，该算法不仅仅能克服非均匀光照的影响，有很好的抗噪效果，还大大减少了分割时间。
침대전통국부역치산법재분할저질량도상시존재분할시간장적문제，제출일충신적국부역치분할산법。해산법수선거제전통역치모형중국부표준차인식，재이용적분도상쾌속계산국부균치，연후이용구득적국부균치화상소점적회도치래계산국부역치。유우강저국부역치계산적시간복잡도，종이감소료도상분할시간。실험결과표명，해산법불부부능극복비균균광조적영향，유흔호적항조효과，환대대감소료분할시간。
In view of the problem of traditional local threshold algorithm that it has long segmentation time when segmenting low quality im-age,we propose a new local threshold segmentation algorithm.The algorithm first gets rid of local standard deviation factor in traditional threshold model,and then uses integral image to rapidly calculate the local mean,finally it uses the local mean obtained and the grey value of pixel to work out the local threshold value.The algorithm reduces the time complexity of local threshold calculation,so that reduces the time of image segmentation.As the experimental application results show,the algorithm not only can overcome the influences of non-uniform illu-mination and has very good anti-noise effect,it also greatly reduces the segmentation time.