Image target detection method, device and storage medium

Abstract

The embodiment of the present invention discloses an image target detection method, device and storage medium. The embodiment of the present invention can acquire the image to be detected, and perform scaling processing on the image to be detected according to different resolutions, and obtain multiple scaled images ; Separately screen the areas that meet the target area type from each zoomed image to obtain multiple initial image blocks; divide each initial image block into multiple areas, and obtain the probability that each area belongs to the target area; from each Extract the region whose probability is greater than the preset threshold from the initial image blocks to obtain a plurality of candidate image blocks; map the plurality of candidate image blocks to the image to be detected, and according to the plurality of candidate image blocks The coincidence degree among them screens out the area where the candidate image blocks that meet the preset conditions are located, and obtains the target area. This scheme has low demand for computing resources and high detection speed, which improves the efficiency and accuracy of image target detection.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an image target detection method, device and storage medium. Background technique [0002] With the development of science and technology and the rise of deep learning, the technology of recognizing targets in images has become one of the most important technologies in computer vision, and the application of deep learning in the field of image target detection has made great breakthroughs. A series of Image object learning methods for deep learning algorithms are proposed. For example, deep learning algorithms such as Faster-RCNN (Faster-Regions with Convolutional Neural Networks features) and YOLO (You Only Look Once). Through these deep learning algorithms, the area where an object is located can be identified from a given image, such as objects such as people, cars, or houses are identified on the image. [0003] At present, since the technical focus of imag...

AI Technical Summary

Problems solved by technology

[0003] At present, since the technical focus of image target detection algorithms based on deep learning such as Faster-RCNN and YOLO is on the accuracy of detection results, the running speed of existing image target detection algorithms does not meet the requirements of actual scenarios. The models of some relatively efficient image object detection systems are very large, which makes the existing image object detection systems run slowly and cannot be implemented on mobile terminals with small computing resource allocation, that is, existing image object detection algorithms are ubiquitous Not only the detection efficiency is low due to the problems of large model and slow speed, but also it is difficult to meet the needs of mobile terminals for real-time detection

Images