论文部分内容阅读
针对具有背景干扰、信噪比低的红外图像,提出了一种基于帧差法和自适应区域生长的红外运动目标检测方法。首先对红外图像进行了高帽变换,以抑制大面积背景的干扰,相邻帧图像间做帧差,初步提取目标区域;其次分析了红外目标的特性,针对其特性提出了一种基于灰度等级的自适应阈值分割方法;最后以帧差法检测的目标质心为种子点,以自适应阈值为分割准则,在预处理后的图像中进行区域生长,最终实现了红外运动目标的检测。结果表明,所提算法可抑制大面积背景的干扰,实现单个和多个红外运动目标的完整提取和检测。
Aimed at the infrared images with low background noise and low SNR, an infrared moving target detection method based on frame difference method and adaptive region growing is proposed. Firstly, the high-hat transform of the infrared image was carried out to suppress the interference of the large-area background, the frame difference was made between the adjacent frame images, and the target area was initially extracted. Secondly, the characteristics of the infrared target were analyzed, Level adaptive threshold segmentation method. Finally, the target centroid detected by the frame difference method is a seed point, and the adaptive threshold is used as the segmentation criterion to grow the region in the preprocessed image. Finally, the detection of the infrared moving target is achieved. The results show that the proposed algorithm can restrain the interference of large area background and achieve the complete extraction and detection of single and multiple infrared moving targets.