Image quality evaluation method and system based on general vision pre-training model

Abstract

The invention belongs to the field of machine learning and data quality evaluation, and specifically relates to an image quality evaluation method and system based on a general-purpose visual pre-training model, aiming to solve the problem of video surveillance image quality in the absence of a large number of manual screening in the prior art. And the problem of low efficiency and accuracy of video quality evaluation. The present invention includes: obtaining an image sample set with a set type of abnormality, performing abnormal screening and labeling; performing sample collection model training through the first training sample set after screening and labeling; setting the collection of the trained sample collection model Threshold, and online collection of the second training sample set with video frame image quality evaluation label; Carry out image quality abnormal classification model training through the second training sample set; Online video frame image and / or through the trained image quality abnormal classification model Classification of quality anomalies in single-frame images. The invention has less manual intervention, and has high efficiency, accuracy and precision of video monitoring image and video quality evaluation.

Description

technical field [0001] The invention belongs to the field of machine learning and data quality evaluation, and in particular relates to an image quality evaluation method and system based on a general visual pre-training model. Background technique [0002] Video surveillance is an important part of the security system. It is widely used in many fields, such as security, finance, transportation, campus security and smart home, because of its intuition, accuracy, timeliness and rich content. The rapid development of computer technology, big data and artificial intelligence provides a larger display stage for the application of video surveillance. At the same time, in the field of security monitoring, the number of cameras is increasing rapidly by 20% per year, which puts forward higher requirements for our management work. Require. [0003] The earliest image and video quality evaluation is usually through personnel to check whether the cameras in use are abnormal. However, ...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the above-mentioned problems in the prior art, that is, in the absence of a large number of manual screenings in the prior art, the efficiency and accuracy of the image and video quality evaluation of video surveillance are low, the present invention provides a method based on general visual prediction An image quality evaluation method for training a model, the method comprising:

Images