论文部分内容阅读
野外复杂背景下红外图像序列目标检测是红外野外监视中的重点问题。大量的背景景物增加了目标检测的难度。文中针对野外复杂背景下红外图像序列的特点,提出了一种实用的运动目标检测算法。该算法包括两个处理步骤:首先,在场景配准后利用帧间差图像提取目标的运动信息,并据此进行目标的粗检测;其次,结合目标运动在时间和空间上的相关性进行精检测。粗检测的低漏判率和精检测的低误判率保证了算法的可靠性。在检测的同时算法确定了目标区域的位置。文中给出了有关的实验结果。
The target detection of infrared image sequence in complex field in the wild is the key issue in infrared field surveillance. A large number of background objects increase the difficulty of target detection. In this paper, aiming at the characteristics of infrared image sequences in the field with complex background, a practical moving object detection algorithm is proposed. The algorithm includes two processing steps: firstly, the motion information of the target is extracted by using the inter-frame difference image after the scene is registered, and then the coarse detection of the target is carried out; Secondly, the time and space correlation of the target motion is refined Detection. The low detection rate of rough detection and the low false detection rate of fine detection ensure the reliability of the algorithm. While testing, the algorithm determines the location of the target area. The paper gives the relevant experimental results.