Infrared image-based small target detection method

Abstract

The invention provides an infrared image-based small target detection method. The method comprises the steps of selecting out a frame image from a group of sequence images by a detection apparatus, and preprocessing the selected frame image to obtain a preprocessed image fD; moving a sliding window to convert the preprocessed image fD into a plurality of independent block images, wherein a matrix D represents the independent block images; performing iterative convergence operation on the matrix D by utilizing a near-end accelerating gradient method to obtain a sparse matrix and a low-rank matrix, wherein the sparse matrix is a target matrix T, and the low-rank matrix is a background matrix B; restoring the background matrix B and the target matrix T to a background image fB and a target image fT respectively through a filter, wherein the preprocessed image fD contains the background image fB and the target image fT; and according to an adaptive threshold, performing threshold processing on the background image fB and the target image fT by the detection apparatus, thereby removing a false alarm. The effective separation of the background image fB and the target image fT in the original image is realized; the specific position of a small target is obtained; and the detection of the small target is realized.

Description

technical field [0001] The invention relates to the technical field of remote sensing, in particular to a method for detecting small targets based on infrared images. Background technique [0002] In recent years, the use of infrared imaging technology to achieve target detection is one of the main technical development directions of modern military weapons and equipment, and it is also one of the important symbols of the automation, intelligence and modernization of military weapon systems. Therefore, scholars from many scientific research institutions at home and abroad Has been committed to the research of this technology. Because the infrared sensor is affected by factors such as the atmosphere, sea surface radiation, operating distance, and detector noise, the size of the long-distance target is small on the infrared image, and even appears as a point. In addition, the signal-to-noise ratio of the image is low, and the background is usually complex, and the target is e...

AI Technical Summary

Problems solved by technology

The background intensity is generally very high, and the noise in the receiver is also very strong. In the case of low signal-to-noise ratio, the intensity of the target is relatively low, and it is often submerged in the strong background noise; at the same time, valuable targets often occupy only A few pixels, lack of geometric shape, texture structure and other features, little information available for detection and recognition system

In summary, it is difficult to detect small targets in infrared images under complex backgrounds.

Images