Target recognition and guidance method of strapdown image guided aircraft

Abstract

The invention discloses a target recognition and guidance method for a strapdown image guided aircraft, and the method comprises the steps of carrying out the feature extraction of a target photo image and a strapdown image, inputting an extracted feature map into a deep neural network, carrying out the recognition of a target region, and obtaining the target position information, therefore, the purpose of inputting a target picture given by an operator into the calculation model and controlling the model to identify and track the target is achieved, further, the model is applied to the strapdown visible light seeker image, the target can be accurately identified and tracked, and uncertainty caused by a complex background and strapdown seeker shaking can be shielded.

Description

technical field [0001] The invention relates to the field of aircraft guidance and control, in particular to a target identification and guidance method for a strapdown image guidance aircraft. Background technique [0002] At present, the application of image guidance in the field of aircraft control is receiving more and more attention. Among them, the strapdown image guidance needs to obtain real-time photos of the field of view through the camera of the strapdown on the aircraft. The field of view is generally large and contains targets. Then further lock the target and know the target position, so as to control the aircraft to fly to the target. [0003] However, the overall effect of the current strapdown image guidance scheme is not ideal, and there are still some problems that need to be further overcome. First, the photos in the field of view are large, and the background area accounts for a very high proportion. Under different application occasions, the background...

AI Technical Summary

Problems solved by technology

[0003] However, the overall effect of the current strapdown image guidance scheme is not ideal, and there are still some problems that need to be further overcome. First, the photos in the field of view are large, and the background area accounts for a very high proportion. Under different application occasions, the background area is very different. It will cause huge interference to target recognition, so in the existing guidance and control schemes, the situation of finding the wrong target or not finding the target occurs from time to time; because the strapdown image will be affected by the vibration of the aircraft, its image noise is large, and it will also Makes it more difficult to identify and filter out targets from field-of-view photographs

In addition, and more importantly, due to the large field of view photos, which may contain multiple targets that can be recognized by the model, it is difficult for the operator and the aircraft itself to select the expected target from them in real time, which leads to the fact that even if the aircraft hits, It may not be able to achieve the expected effect. In fact, such a situation is more likely to cause the aircraft to miss the target

Images