Video image fuzzy anomaly detection method based on machine learning

Abstract

The invention discloses a video image fuzzy anomaly detection method based on machine learning and belongs to the technical field of intelligent traffic. The method comprises the following steps of: firstly converting high-definition video color images into grayscale images; manually classifying the converted grayscale images into two categories according to fuzziness and non-fuzziness and calculating the gradient histogram feature of each image; using the gradient histogram features as the classification features, training the classification features by using the support vector machine, and saving the trained parameters; and using the trained parameters to calculate the gradient histogram feature of newly input images and then to calculate the output result of the support vector machine; and judging the image as fuzzy image if the support vector machine is positive or judging the image as a non-fuzzy image if the support vector machine is negative. The video image fuzzy anomaly detection method based on machine learning has strong robustness and can be applied to determining whether the video image has fuzzy problems or not.

Description

technical field [0001] The invention relates to the technical field of intelligent transportation, in particular to a method for detecting blurred anomalies of video images based on machine learning. Background technique: [0002] Since there are a large number of surveillance cameras in the intelligent transportation field of each city (more than 1,000 in each city), it is time-consuming and labor-intensive to rely on manual methods to troubleshoot cameras with fuzzy problems, and it is very prone to omissions or negligence, which makes the cameras that should be maintained If it is not detected, the camera will often blur due to focusing problems or equipment aging. This kind of blurring is usually not easy to detect, but it will bring disastrous consequences for later analysis of video surveillance or use of surveillance video to restore the scene . Contents of the invention [0003] In view of the above problems, the technical problem to be solved by the present inven...

AI Technical Summary

Problems solved by technology

[0002] Since there are a large number of surveillance cameras in the intelligent transportation field of each city (more than 1,000 in each city), it is time-consuming and labor-intensive to rely on manual methods to troubleshoot cameras with fuzzy problems, and it is very prone to omissions or negligence, which makes the cameras that should be maintained If it is not detected, the camera will often blur due to focusing problems or equipment aging. This kind of blurring is usually not easy to detect, but it will bring disastrous consequences for later analysis of video surveillance or use of surveillance video to restore the scene

Images