Low-contrast images have the characteristics of high spatial correlation of adjacent pixels and no obvious gray change,so it is very difficult to segment small target from low-contrast image. A new low-contrast image segmentation algorithm to extract small targets was presented. Spatial distribution of pixels was obtained based on local uniformity and local entropy,and low-contrast images were segmented by using the gray and spatial properties of target images. Experimental results show that the algorithm can effectively achieve image segmentation of low contrast small target. Comparing with other algorithms,the algorithm has higher segmentation accuracy and strong noise resistance,and it can effectively reduce the excessive segmentation.
参考文献
相似文献
引证文献
引用本文
曾毅,郭龙源,赵娇燕.基于邻域空间特征的低对比度小目标分割算法[J].激光与红外,2017,47(4):507~512 ZENG Yi, GUO Long-yuan, ZHAO Jiao-yan. Low contrast small target segmentation algorithm based on neighborhood space feature[J]. LASER & INFRARED,2017,47(4):507~512