车辆检测方法、装置、电子设备及计算机可读存储介质
By acquiring key points of the target vehicle during vehicle detection and grouping them using the mass field feature map and the original vector, the low accuracy and false positives/false negatives of existing vehicle detection methods when the image quality is poor are solved, achieving higher detection accuracy.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SF TECH CO LTD
- Filing Date
- 2022-06-29
- Publication Date
- 2026-07-17
AI Technical Summary
Existing vehicle detection methods have low accuracy when image quality is poor, and convolutional neural network detection boxes tend to include redundant scene information from non-vehicle areas, leading to false positives and false negatives.
By acquiring the coordinates of key points of the target vehicle in the image to be detected, the matching point pairs are grouped using the target mass field feature map and the original vector to identify the key points of the same vehicle, and the mass field information is combined to determine whether the matching point pairs are key points of the same vehicle.
It improves the accuracy of vehicle detection, avoids the impact of image quality issues on detection, and reduces false positives and false negatives.
Smart Images
Figure CN117372316B_ABST