【作者】 王新增；

【导师】 黄山；

【作者基本信息】 四川大学 ， 控制理论与控制工程， 2005， 硕士

【摘要】 近年来,随着计算机和数字图像处理技术的飞速发展,红外目标检测技术开始广泛应用于制导、跟踪、自动控制、人工智能等诸多领域。由于红外弱小运动目标具有距离远、面积小、形状特征弱、细节特征部分丧失、信噪比低等特点,对红外弱小运动目标的检测是目标检测的重点和难点。目前,国内对红外弱小运动目标检测方法的研究比较踊跃,发展比较快,具有一定的基础,但要达到满意的技术水平仍存在很大的差距,况且任何检测方法都具有一定的针对性和局限性。 关于红外目标检测理论和方法的研究,始终是为了不断提高低信噪比、复杂背景下的目标检测能力;提高检测算法的探测概率、适用范围、运算效率、可靠性、降低虚警率。所以,对红外弱小运动目标检测方法进行研究仍然是一个具有挑战性的课题。 本文针对某电子靶场试验中获得的空中红外动态序列图像数据,在分析图像特征的基础上,对其中的红外弱小运动目标的检测方法进行了研究。在比较前人研究的基础上,提出了基于灰度形态学的背景抑制方法对图像进行预处理,利用自适应阈值分割方法对预处理后的图像进行阈值分割的帧内检测方法;以及基于图像序列相邻帧中可疑目标之间的位置关系,利用八邻域判决算法对目标进行检测的帧间检测方法。 实验结果表明,本文提出的空中红外弱小目标检测步骤合理;算法切实可行;检测速度快,检测率高。更多还原

【Abstract】 In recent years, With the quick development of computer and digital image processing technology, the infrared target detection technique start to be extensive apply in make to lead, follow, automatic control, artificial intelligence etc. many realm.Because the infrared moving weak and small target has the characteristics that the distance is far, the area is small, the shape characteristic is weak, the detail characteristic part lose and the SNR is low etc., to the infrared moving weak and small target detection is a important and difficult point of the target detection.At present, the research the infrared moving weak and small target detection method is eagerly and quickly at domestic, has the certain foundation, but there is still a huge gap to attain the satisfactory technique level, besides any detection method aimed at some point and has its limitation.The research of the infrared moving weak and small target detection theories and methods, always for raising the SNR, the detection ability of target within complicated background; Increase the detection probability, coverage, efficiency and reliabi1ity of detection algorithm descend falsely rate of alarm. So, It’ s still a challenging subject that the methods about infrared moving weak and small target detection.We used the infrared sequence dynamic image data that acquired at some electronic range, based on analysis of features of image, to research the method of infrared moving weak and small target detection. Based on the previous research, a pretreatment based on greyhound scale morphologic background-constrict; a frame detection method used of self-adapt threshold to segment the pretreated image, a frame detection method used 8 close area based on the relationship between suspicious targets in the image sequence close frames are presented in this paper.The result of experiment show that the arithmetic is reasonably, andthe system is reliable and effective.更多还原