Target detection model training method and target rapid detection method

Abstract

The invention discloses a target detection model training method and a rapid target detection method. The training method comprises the following steps: adding a target region feature enhancement layer during training; performing feature extraction on the training sample by using a feature extraction unit, and averaging channels of a feature map output by a previous-level feature extraction unit to obtain a first feature value matrix with a normalized channel number; traversing each pixel point in the training sample to generate a second eigenvalue matrix; multiplying the element values of thefirst eigenvalue matrix by the element values of the second eigenvalue matrix to obtain a third eigenvalue matrix; multiplying the third eigenvalue matrix by a preset adjustment function, then performing element value addition on each channel eigenmatrix of the eigenmap to obtain a target enhanced eigenmap, and inputting the target enhanced eigenmap into a next eigenextraction unit. According tothe invention, the network is fully trained, the relationship between the network depth and the detection precision is balanced, the background perception capability of the feature map is enhanced, the detection precision is high, the calculation is simple, and hardware platform transplantation is facilitated.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and more specifically relates to a training method of a target detection model and a fast target detection method. Background technique [0002] In the land battlefield, tanks are the main combat tools. In order to prevent enemy attacks, ambushes, and protect our important personnel and equipment, it is very necessary to detect and alert vehicles such as tanks. The battlefield alert of mobile platform detection and identification equipment can realize all-round reconnaissance and become more intelligent. At present, the existing traditional detection algorithms are computationally complex and difficult to adapt to the complex battlefield environment, while the general deep learning detection algorithms are computationally complex and difficult to achieve real-time detection on mobile platforms. [0003] The following difficulties exist in the detection and recognition of moving platform ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at at least one defect or improvement requirement of the prior art, the present invention provides a training method of a target detection model and a fast target detection method, using a method of target area feature enhancement to fully train the network; designing a half-channel feature information integration structure to Enhance the distinguishability of eigenvalues ​​for targets and backgrounds; integrate low-dimensional feature maps with high-dimensional feature maps to perform detection at different scales, and improve target detection accuracy at various scales; balance the relationship between network depth and detection accuracy, and This method is applied to the detection of tanks and vehicles, thereby solving the problem of certain limitations in the application of tanks and vehicles detection algorithms in the prior art to moving platforms in complex battlefield backgrounds

Images