Infrared weak and small target real-time detection method and system based on target detection model

Abstract

The invention provides an infrared weak and small target real-time detection method and system based on a target detection model, and the method comprises the steps: constructing an infrared weak and small aircraft detection data set, and carrying out the redesign of a feature extraction backbone network, so as to enable the network to be more suitable for the feature extraction of a weak and small target; a spatial pyramid pooling layer is added behind the backbone network again to obtain different receptive fields, so that the positioning capability of the network model to weak and small targets is enhanced; and finally, aiming at the relatively fixed and single size of the infrared weak and small target, reducing detection heads, and fusing shallow layer information and deep layer information through a characteristic pyramid tower, so that the network has higher detection accuracy and stronger robustness. The weak and small target detection accuracy and speed under the complex background are improved, and the robustness and adaptability of a weak and small target detection algorithm are improved.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a real-time detection method and system for infrared weak and small targets based on a target detection model. Background technique [0002] Infrared weak and small target detection is a research difficulty and focus in the field of target detection, and weak and small target detection plays an important role in national defense and military. In the past, infrared weak and small target detection mainly relied on the method of searching the region of interest based on the characteristics of the target, the threshold segmentation method using the physical characteristics of infrared imaging, and the background prediction modeling method. However, these algorithms have low detection efficiency and are difficult to deal with complex and changeable. Scenes. In recent years, with the development of deep learning, target detection algorithms based on deep learning have made g...

AI Technical Summary

Problems solved by technology

At present, most networks only pay attention to the performance of the network, and continuously increase the complexity of the network to improve the detection effect of infrared weak and small targets. The characteristics of infrared weak and small targets are lost, so it is of great significance to design a dedicated network for the characteristics of infrared weak and small targets, so that it can run in real time on resource-constrained devices

Images