超大视场红外凝视成像系统具有视场大、被动探测、凝视探测等独特优势，但当系统用于弱小目标检测时，由于背景复杂、噪声干扰、目标信息少等问题，检测的准确性和效率往往不高。本文通过采用超大视场中空时域融合处理的思想，提出了一种基于最大化背景模型进行背景抑制的改进方法。该方法首先通过图像预处理、多帧差分选取研究区域；然后通过改进的背景预测模型检测疑似目标点；最后，利用邻域相关准则判定真实目标。通过实验证明：该方法将原方法中的目标信噪比提高了5倍以上，灰度值提高了10倍以上，而且保留了更多的目标信息。同时目标检测时间减少了50%以上，提高了检测准确性和检测效率。%The super wide-field infrared staring system has unique characters such as vast detection airspace, passive detection, staring detection, and so on. However, when a super wide-field infrared staring system is used to detect weak and small targets, it may show complex background, more noise jamming and little target information. So the detection accuracy is always low. By using the idea of spatial-temporal fusion processing in super wide-field image, an improved method based on maximum background model prediction is proposed in the paper. Firstly, image preprocessing and multi-frame difference are carried out to select study regions. Then, the improved model is used to detect all suspicious targets in whole infrared image. At last, the real target is confirmed by neighborhood correlation rule. The experiment proves that:the Signal to Noise Ratio(SNR) of proposed method is increased more than5 times and the gray value of target is increased more than 10 times. Meanwhile, more target information is preserved than before. The time of target detection is decreased by more than 50% and the accuracy and efficiency of detection are improved.
展开▼
