Target detection method and device, computer device and computer readable storage medium

Abstract

The invention provides a target detection method. The target detection method comprises the following steps: acquiring a training sample set, wherein the training sample set comprises a plurality of target images labeled with target positions and target angle types; training an accelerated regional convolutional neural network model by using the training sample set to obtain a trained acceleratedregional convolutional neural network model; acquiring a to-be-detected image; performing target detection on the to-be-detected image by using the trained accelerated regional convolutional neural network model to obtain a target region of the to-be-detected image and the target angle type of the target region. The invention further provides a target detection device, a computer device and a readable storage medium. Through the target detection method, the target detection device, the computer device and the readable storage medium, rapid target detection with high detection rate can be achieved.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a target detection method and device, a computer device and a computer-readable storage medium. Background technique [0002] Existing object detection techniques include object detection based on simple pixel features or hand-designed complex features. Using simple pixel features, such as representative HAAR, pixel difference, etc., although the calculation efficiency is high and the real-time performance is good, but the robustness to factors such as complex and diverse background changes is poor, and the detection accuracy is lacking. However, based on complex features designed by hand, such as HOG in DPM, although the feature expression is better and the robustness is stronger, it is difficult to meet the real-time requirements because GPU acceleration cannot be used, and the calculation is complicated on the CPU. [0003] Existing object detection technique...

AI Technical Summary

Problems solved by technology

Although GPU computing solves the computational problem of extracting convolutional features, candidate region extraction still takes a considerable amount of time

In addition, since the whole scheme is a framework process of extracting candidate regions first and then classifying them, end-to-end detection cannot be realized, and the application is relatively cumbersome

[0004] In addition, due to different shooting angles, the appearance of the target object will change greatly on the image. The existing target detection technology does not consider the problem of shooting angle, resulting in a low target detection rate.

Images