An infrared target detection algorithm based on region growing was presented to solve the problems of low contrast, complex background, and unconductive to objective testing. First, on the basis of morphological filtering mechanism, and using overall distribution of the texture image for progressive scanning, effectively selecting maxima minimum point, the size of the structural elements was determined adaptively for the image filtering. Second, based on the automatic extraction of the seed region, growth threshold was sentenced by dynamically adjusting the growth conditions, determining the target to be detected . Experimental results show that the proposed method compared with the NCC methods and OTSU method has the advantages of 10% for detecting accuracy rate and higher precision.%针对前视红外图像对比度低、背景干扰严重、不利于目标检测等问题，提出了一种基于区域生长的目标检测方法。首先，在形态学滤波机理的基础上，利用图像的整体纹理分布特点进行逐行扫描，选取有效的极大值点、极小值点，自适应确定结构元素的大小进行图像滤波。其次，在自动种子区域提取的基础上，通过生长条件的判决准则动态调整生长阈值，确定待检测目标。实验结果表明：该方法与NCC方法和OTSU方法相比，检测准确率提高了10%，精确度较高。
展开▼
