针对红外图像中背景与目标的复杂性和多样性给弱目标检测带来的困难,提出基于局部方差的二维最大熵分割与遗传算法相结合的阈值分割方法.本文对变电站绝缘套管的红外图像进行了处理,首先采用形态学顶帽变换的方法对其进行增强,然后对图像进行局部方差映射并建立二维直方图,最后结合遗传算法计算出二维最大熵分割阈值进而实现分割.实验结果表明,该方法改善了红外弱目标的提取效果,大大提高了阈值计算的效率,待提取目标越小,遗传算法的作用越能得到体现.%The complexity and diversity of the background and targets in the infrared images brought difficulties to the weak target detection. In order to overcome this deficiency, a 2-D maximum entropy threshold segmentation method based on local variance is proposed, which is combined with genetic algorithm. In this paper, the infrared image of Substation Insulation casing is processed. At first, the image is enhanced by using the method of morphological Tophat transform. Then, Local variance mapping for image is made and 2-D histogram is established. Finally the combination of 2-D maximum entropy and genetic algorithm is used to calculate the optimal segmentation threshold and actualize threshold segmentation. The experimental results show that this algorithm can improve the extraction effect of infrared weak target and greatly improves the efficiency of threshold calculation. The smaller the target is, the greater role the genetic algorithm plays.
展开▼
