Object state detection method and device

Abstract

The embodiment of the invention provides an object state detection method and device. The method comprises the following steps; acquiring a to-be-detected image aiming at a preset detection scene, extracting a movement foreground region from the to-be-detected image and extracting a movement foreground object from the movement foreground region; acquiring the first actual height H1 of the movement foreground object and judging whether the first actual height H1 is higher than the second actual height H2 of a preset reference plane; if the first actual height H1 is higher than the second actual height H2, judging that the movement foreground object is in the standing state when the first actual height H1 is higher than the second actual height H2 and the width of the movement foreground object satisfies the preset condition. Visibly, the actual height of the movement foreground object is acquired by using the depth information thereof through the scheme provided by the embodiment of the invention, and whether the movement foreground object is in the standing state is judged according to the actual height, and the standing state detection is realized without needing the detection on the stand-up action, the missing detection situation caused by short action duration time is avoided, and the accuracy rate of the object state detection is improved.

Description

technical field [0001] The invention relates to the technical field of video surveillance, in particular to a method and device for detecting an object state. Background technique [0002] In practical applications, it is often necessary to video record a certain scene (teaching, meeting, etc.). For example, in the process of intelligent teaching, it is often necessary to video record the teaching scene. Give close shots. In order to realize the close-up of the close-up picture of the students standing up to speak, it is first necessary to be able to detect the position of the students standing up to speak in the teaching scene. [0003] When detecting the standing state of students, such a detection scheme is provided in the prior art: first, use the main camera to obtain continuous N frames of images for the teaching scene, and determine the student’s status according to the change trend of the pixels in the continuous N frames of images Direction of movement (such as up...

AI Technical Summary

Problems solved by technology

[0004] Although the above scheme can realize the detection of the standing state of the students, the above scheme needs to determine the movement direction of the students according to the change trend of the pixels in the continuous N frames of images, that is, it is necessary to detect the student’s movement direction according to the change trend between consecutive images. Whether there is an action of rising (possibly standing up); however, the duration of the student standing up is usually very short, such as 1 second, etc., so it is easy to appear that the student did rise up but failed to shoot due to the short duration The situation of seeing the image caused the detection of students standing up to be missed, thus reducing the accuracy of the detection of students' standing up state

Images