A Method of Obtaining Abnormal Areas Using Parallax Information and 3D Information

Abstract

The invention discloses a method for obtaining abnormal areas by using parallax information and three-dimensional information, using the changes of parallax values ​​and three-dimensional coordinates before and after the occurrence of events to detect whether there are abnormal points, and adopting the mean value matrix E of the parallax image to represent the parallax without abnormal conditions Value, and then use to create a new image with the same size as the parallax image, take part of the data to calculate the variance and mean value, and then use the difference between the pixel value in the pixel queue corresponding to the current point and the mean value of the new image part data to compare with the set threshold value, Judging the points of abnormal areas; this technical means can overcome the disadvantages of pre-setting standard images, overcome the influence caused by the existence of light, buildings, plants, and other disturbing objects caused by time changes, and use the parallax before and after the event value to determine the outlier point. In addition, the present invention can finally mark different types of abnormal points according to the attributes of the abnormal points with different characteristics, such as marking settlement points, foreign matter points, and interference points.

Description

technical field [0001] The invention relates to a visual image detection data processing technology, in particular to a method for acquiring abnormal regions by using parallax information and three-dimensional information. Background technique [0002] In the visual image detection data processing technology, the extraction of image information often adopts a relatively simple means to meet the expected technical requirements, such as face recognition, and image information is often used in road traffic, such as the technology to identify abnormal areas of images In the process, the standard image information is often collected in advance, and then in the later image acquisition process, the difference between the later image and the standard image is used for differential comparison to find the difference point and finally output the difference point. The disadvantage of this method is that it requires high quality, High-standard standard images, but in the actual applicati...

AI Technical Summary

Problems solved by technology

[0002] In the visual image detection data processing technology, the extraction of image information often adopts a relatively simple means to meet the expected technical requirements, such as face recognition, and image information is often used in road traffic, such as the technology to identify abnormal areas of images In the process, the standard image information is often collected in advance, and then in the later image acquisition process, the difference between the later image and the standard image is used for differential comparison to find the difference point and finally output the difference point. The disadvantage of this method is that it requires high quality, High-standard standard images, but in the actual application process, because the standard images are collected on-site, the standard images will be inaccurate due to external changes in light, buildings, and plants, often causing false positives for differences, such as noon time The difference between the standard image collected in the evening and the later image collected in the evening is not only the image itself, but also the huge difference in light. There is a huge difference, so if you keep using the previous standard image for comparative analysis, it will cause a lot of false positives

Therefore, this technique of comparing and analyzing standard images collected on-site with later images cannot be well applied and promoted.

Images