Target capturing method and system of edge device and related device

Abstract

The invention is suitable for the field of artificial intelligence application, and provides a target capturing method and system of edge equipment and related equipment. The method comprises the steps: obtaining video data, and putting frame images obtained through frame splitting according to a time sequence into a detection queue; performing target detection on each frame of image in the detection queue by using a deep learning algorithm, and marking a detected face frame, a human shape frame, a vehicle frame and a license plate frame; target tracking is carried out on each face frame, each human-shaped frame and each vehicle frame, the same frames of target tracking results are endowed with the same tracking ID, and the results are output; performing one-way binding association on the face frame and the humanoid frame in each frame of image, and performing two-way binding association on the vehicle frame and the license plate frame; image judgment is carried out on the frames with the same tracking ID, and a face frame, a human shape frame, a vehicle frame and a license plate frame are graded and output. According to the invention, the resource occupancy rate is reduced, and the problem of insufficient performance of edge equipment is solved.

Description

technical field [0001] The invention belongs to the field of artificial intelligence applications, and in particular relates to an edge device target acquisition method, system and related devices. Background technique [0002] With the rapid spread of the practical application of the smart city security system, it has become a technical difficulty for the security system to extract high-value information at high speed and effectively from the massive monitoring image data obtained from the edge monitoring equipment. For edge devices, the disadvantage of weak computing power has always been the bottleneck of their development. Although the performance of hardware devices has improved significantly in recent years, it is still difficult to meet the needs of computing-intensive applications such as real-time image processing, especially for applications in specific fields. The response speed, stability and reliability of the system have high requirements, which has caused more...

AI Technical Summary

Problems solved by technology

[0004] Embodiments of the present invention provide a target acquisition method, system and related equipment for edge devices, aiming to solve the problem that the integration difficulty of existing artificial intelligence algorithms deployed on edge devices is high, and the computing performance is limited by the computing power and foundation of edge devices. equipment, issues that cause slow processing

Images