An infrared small target detection method based on multi-frame regression deep network

Abstract

This project belongs to the field of target detection in infrared remote sensing image processing. It provides a small infrared target detection method based on multi-frame regression deep network, and solves the problems of low robustness and narrow application range of existing detection methods based on single-frame images. There are a large number of false detection problems in infrared small target detection. Its main scheme includes creating a multi-frame infrared small target data set, obtaining a single-frame target candidate area based on the difference union map extracted from multiple frames and local variable threshold segmentation, and extracting the one-dimensional feature of the candidate area through multi-frame trajectory association. and create a feature dataset. Input the data set into the long-term and short-term memory regression network for training, input the test data to the regression network, and obtain the data category according to the network output. target detection result.

Description

technical field [0001] An infrared small target detection method based on a multi-frame regression depth network is used for infrared small target detection in infrared remote sensing images, and belongs to the field of target detection in remote sensing and infrared image processing. Background technique [0002] Infrared search and track (IRST) system has extremely high military value. As a basic function of IRST system, infrared small target detection technology is indispensable in infrared search, infrared early warning and infrared tracking. Due to infrared imaging conditions, noise or interference is inevitable in infrared images. Among them, the false alarm source is similar to the target in the satellite infrared image, and both have high grayscale, so it may cause the false alarm of the remote sensing early warning system. Due to the difficulty of detecting small infrared targets, although scholars at home and abroad have proposed a variety of detection algorithms,...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is: the present invention provides a method for infrared small target detection based on multi-frame regression depth network, which overcomes the fact that the existing method cannot realize real-time accurate positioning in the face of the trade-off between detection accuracy and speed of infrared image small targets The problem of small target position can improve the detection ability and reduce the false alarm rate in the case of real-time detection

Images