Infrared weak and small target detection method under cloudy background based on discrete fractional Brownian random field

Abstract

The present invention proposes a method for detecting small and small infrared targets under cloudy background based on discrete fractional Brownian random fields. Firstly, the infrared image is preprocessed by using morphology and Gaussian filtering; secondly, the multi-scale Hurst exponent map is calculated for the preprocessed results; Then, by comparing the Hurst exponents at different scales, and inverting and enhancing the saliency of the image, the final Hurst exponent map is obtained; then the improved inter-class variance is used to further remove noise and enhance the saliency of the target; finally find in the figure The point with the largest gray value is the location of the weak infrared target. The invention utilizes the characteristic that the Hurst parameter can characterize the self-similarity (that is, the uniformity of the gray surface) of the same image area, and combines multi-scales to improve the detection robustness.

Description

technical field [0001] The invention belongs to the technical field of image processing, and specifically relates to a method for detecting small infrared targets in a cloudy background based on a discrete fractional Brownian random field. Background technique [0002] With the continuous enhancement of national defense scientific research strength, in the field of national defense such as infrared guidance, space machine early warning and target monitoring, the detection technology of weak and small targets in infrared images has always been a focus of attention and research by scholars at home and abroad. However, due to the long imaging distance of infrared images, the large background noise, the complex imaging environment, and the uncertain trajectory of the target, this research has always been a difficulty. [0003] Fractional Brownian Motion (FBM) was first proposed by H.B.Mandelbrot in 1965. It is used to simulate various noises with fractal characteristics and can ...

AI Technical Summary

Problems solved by technology

At present, in the field of infrared small target detection, most of them use the fractal dimension in fractal theory to detect targets. Commonly used algorithms include box counting method, curve fitting, etc. These algorithms have high complexity and long running time. Infrared small target detection under low detection probability is low, false alarms are prone to occur, and the robustness is not strong

Images