RX infrared small target detection algorithm based on multi-scale local strength and gradient fusion

Abstract

Infrared small target detection under complex background is still a challenging computer vision task. An image usually contains complex background, weak targets and chaotic noise. In order to solve this problem, we must try our best to enhance the target signal to noise ratio (SNR) while suppressing the complex background factors. We proposed an RX infrared small target detection algorithm by analyzing multi-scale local gradient and local intensity fusion feature. According to the gradient points to the target center and the intensity value of the target pixel is greater than the intensity value of its local adjacent pixels, We used multi-scale information to calculate the local gradient map and local intensity map of infrared image and used RX global detection method to extract small targets. The experiment shows, our algorithm has good performance on detection result. It has good robustness and effectiveness, which can be applied to infrared small target detection under complex background.