低信噪比单帧图像中的小目标检测方法地研究.pdf

能好，并且能较准确地提取小目标区域的形状特征，实现对弱小目标的检测定位。 关键词：弱小目标，图像滤波，目标检测，小波变换，遗传算法 Abstract With the development of modern warfare technology, long-range reconnaissance and long-distance precise strike have become the important means of modern warfare. Because the optical sensors are passive sensors, their maximum detection range is critical. When the target is far away from the optical sensors, the signal-to-noise ratio of the image generated by them is low, an advantageous position can be occupied if the target is discovered and caught as soon as possible. Therefore, research on the dim small target real-time detection is of great significance. In order to conquer the difficulty of dim small target detection and meet the real-time of algorithm, single-frame image detection methods are used. Through image filtering, the background noise can be reduced and the image’s SNR can be improved, then image segmentation algorithm is used to separate the dim small target and the complex background, so as to achieve target detection. This paper is supported by the innovation special fund for graduate student of Jiangxi Province and innovation project of Nanchang Hangkong University, and committed to research on the methods of dim small target detection based on single-frame images, the main work and results are as follows: （1）The filtering methods which are fitted for low SNR image have been studied. Wavelet transform is used to image denoise because it has good time-frequency localized characteristics so that it can achieve a better balance in the protection of signal feature and restraint of noise. Knows from traditional wavelet de-noising methods that base on magnitude, an algorithm bases on wavelet phase is proposed for noise suppression, the