自动目标检测是图像精确制导技术的重要研究内容。针对地面复杂背景条件下目标难以检测的问题,提出一种结合灰度形态学滤波与直方图独立峰搜索的目标分割方法。首先,对复杂背景下的图像进行形态学滤波和增强,增强目标和背景的灰度对比度;然后利用文中提出的图像直方图独立峰搜索方法确定图像阈值,对图像进行分割。实验表明:文中方法是有效的,对复杂背景下的目标图像可取得较好的分割效果。

Automatic target detection is an important part of the research on imaging precision guided technology. Since it is difficult to recognize targets in a complex background, a segmentation algorithm based on gray-scale morphology filter and searching independent hump of the image histogram has been proposed. Firstly, morphology methods had been used to filtering the image noise and enhance the gray scale contrast between the object and the background; and then determine the histogram threshold by searching independent hump of the image histogram. The new approach is proved to be feasible and effective by some comparative experiments.