Rapid target detection system and electronic equipment

Abstract

The invention relates to a rapid target detection system and electronic equipment. The rapid target detection system comprises a feature extraction model obtained by pre-training, a processor and a memory in which a computer program is stored; anchor point parameters are set in the feature extraction model; the anchor point parameters comprise anchor point basic sizes, a group A proportional relation value and a group B scaling relation value; the feature extraction model comprises B feature extraction modules, the scaling relation value corresponding to the bth feature extraction module is Qb, and the bth feature extraction module comprises N channels {Tb1, Tb2,... TbN}. According to the system, the calculation amount required by target detection is reduced, the target detection efficiency is improved. The system can be suitable for devices with weak processing performance, such as edge equipment, mobile terminals and embedded equipment.

Description

technical field [0001] The invention relates to the technical field of computers, in particular to a fast target detection system and electronic equipment. Background technique [0002] Object detection is a basic research topic in the field of computer vision. It is the basis of many computer vision problems such as object tracking and face recognition, and has a wide range of application prospects. Target detection means that for any given image, the algorithm detects and recognizes specific objects in the image, such as pedestrians, objects, etc., and returns the position information of the target, which is generally represented by a rectangular box. Target detection itself is a complex problem. The target itself has various state changes. At the same time, according to different imaging conditions, different images such as occlusion, illumination, and posture will be generated, which further increases the complexity of the scene. These problems will eventually lead to t...

AI Technical Summary

Problems solved by technology

However, SSD has at least the following disadvantages: (1) The network backbone (skeleton) part adopts the VGG network architecture, which has a large amount of calculation and is too cumbersome. It is used on devices with weak processing performance such as edge devices, mobile terminals, and embedded devices. , it is often impossible to process the target detection data in real time; (2) when extracting features, the sampling is too dense, but in fact some sampling is unnecessary, which will also increase the calculation amount of target detection

Images