基于邻域灰度熵和分类的红外弱小目标检测

Infrared Dim Target Detection Based on Neighborhood Gray Entropy and Classification

根据红外图像弱小目标和背景特征,提出了基于邻域灰度熵和分类的红外弱小目标检测方法。先对红外小目标图像进行复Contourlet去噪,并设计一滑动窗口扫描整幅图像,将滑动窗口内的像素分为目标和背景两类,利用两类的灰度均值差及窗口中心像素的邻域灰度熵进行背景抑制,再用指数交叉熵阈值法分割出小目标。实验证明该方法简单有效,易于硬件实现。

According to the characteristics of small target and background in infrared images, an infrared dim target detection method based on neighborhood gray entropy and classification was proposed in this paper. The infrared image of dim target was denoised by complex contourlet. A sliding window was designed to scan the whole image. The pixels in the sliding window were divided into two categories as object and background. The difference of grayscale means of two categories and the neighborhood gray entropy of central pixel in the window were used to suppress the background. The exponential cross entropy thresholding method was used to detect the dim target. The experiment results showed that the method was simple, effective and easy to implement by hardware.