Improved Densebox target detection method and device and storage medium

Abstract

The invention discloses an improved Densebox target detection method, an improved Densebox target detection device and a storage medium, and aims to solve the technical problems of large network parameter redundancy, high calculation complexity and insufficient target detection precision of an existing target detection algorithm. The method comprises the following steps: acquiring historical roadimage data, and carrying out image preprocessing; performing central region adaptive perception based on the preprocessed image to obtain a training sample set of the road image; training a pre-constructed CNN network by using the training sample set to obtain a trained CNN network model; and processing the to-be-detected road image by using the trained CNN network model to obtain a road target detection result. The invention optimizes the CNN network structure, gives full play to the performance of the network, effectively balances the back-propagation loss weighting of various scales of targets through employing the adaptive sensing center region, and improves the target detection precision.

Description

technical field [0001] The invention relates to an improved Densebox target detection method, device and storage medium, belonging to the technical field of ADAS. Background technique [0002] With the development of urban traffic and the substantial increase in car ownership, road traffic safety has become the focus of the whole society. At the same time, car safety has become a global social problem. All countries in the world try their best to adopt various digital electronic technologies and methods to reduce traffic accidents and casualties, and at the same time improve the active safety performance of automobiles. In terms of automotive active safety technology, target detection is the basis and key technology of ADAS technology. The data of the road ahead is obtained in real time through the camera, and the road conditions around the vehicle are monitored in real time. When danger comes, the driver can be reminded in time or directly. The vehicle should intervene rea...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problems of large network parameter redundancy, high computational complexity and insufficient target detection accuracy in the Densebox target detection algorithm in the prior art, the present invention proposes an improved Densebox target detection method, device and storage medium. The self-adaptive perception center area of ​​the label box is adapted to targets of different scales, improving the recall rate and target detection accuracy of medium and long-distance small targets, and at the same time modifying the original network to improve network utilization

Images