摘要
提出了基于复杂低空背景条件下新的图像序列跟踪方法。首先在全视场范围内采用自适应滤波法搜索目标 ,连续几帧检测到目标后 ,进入小视场范围内分割检测目标进行精确跟踪。若目标丢失 ,再返回全视场模式搜索目标。试验结果表明 ,在背景噪声较为强烈的情况下 ,该方法依然能有效地检测跟踪目标。由于采取了小视场跟踪的策略 ,减少了计算时间 ,实时性较好。
A new detection method is suggested for low altitude infrared small targets based on complex background. The method is to suppress the noise and smooth the image in the entire view using median filters, then detect the target with adaptive filters. After the target is detected for several continuous frames, accurate track begins in a small view using grads-reciprocal filter and adaptive filter. We will detect again the target in large view once we lose it in the small view. Tests show that even in strong noise the target will be detected efficiently with this method, and the calculating time is greatly deduced.
出处
《电子工程师》
2004年第9期40-42,共3页
Electronic Engineer
关键词
红外图像
背景抑制
自适应滤波
目标检测
IR image, noise suppression, adaptive filtering, target detection
