Snapshot optimization method and device based on target features, computer equipment and medium

Abstract

The embodiment of the invention discloses a snapshot optimization method and device based on target features, computer equipment and a medium. The method comprises the following steps: creating a feature linked list and a picture linked list; generating a target event according to the detected target; extracting a target feature to obtain a feature vector of the target image; adding the feature vector to a feature chain table for caching; cutting out a target local picture; adding the target local picture to a picture linked list for caching, and starting a preferred timer of the target event; and judging whether the optimal timer is overtime or not, and if the optimal timer is overtime, respectively extracting an optimal feature vector and a corresponding target local picture from the feature linked list and the picture linked list according to the confidence coefficient. According to the invention, target feature extraction is carried out on the target image through the pre-trained target recognition model, and caching and comparison are carried out on the target features and the target local image, so that the target image optimization effect is improved.

Description

technical field [0001] The present invention relates to the field of image processing, more specifically to a method, device, computer equipment and media for capturing and optimizing based on target features. Background technique [0002] With the development of technology and the improvement of image processing requirements, in an application scenario, the amount of image data that can be acquired by a shooting device increases geometrically. At present, most of the shooting equipments simply store the collected image data, which leads to the fact that there are many low-quality image data in the process of image data application, resulting in poor data integrity of the collected image data. Feature availability is low. [0003] At present, although it is also used in the image optimization method, the current image optimization is only judged by the confidence of the target detection, and the effect of deduplication is not good. Contents of the invention [0004] The ...

AI Technical Summary

Problems solved by technology

At present, most of the shooting equipments simply store the collected image data, which leads to the fact that there are many low-quality image data in the process of image data application, resulting in poor data integrity of the collected image data. low feature availability

[0003] At present, although it is also used in the image optimization method, the current image optimization is only judged by the confidence of the target detection, and the deduplication effect is not good.

Images