Object motion trajectory recognition method, device, equipment and storage medium

Abstract

The embodiment of the invention discloses a method, apparatus, device and storage medium for recognizing object motion trajectory. The method comprises the following steps of: acquiring an image of atarget object acquired by an image collector, wherein at least three feature points located on different straight lines are calibrated in an acquisition area of the image collector; The actual position of the target object is determined based on the image positions of the target object and at least three feature points in the image of the target object and the actual positions of the at least three feature points. An embodiment of the present invention determines an actual geographic position of a target object by using a relative positional relationship between the target object and an imageposition of each feature point of which the actual geographic position is known, There is no need to establish the transformation relation of the position coordinates repeatedly, which has strong generality and great practical significance, and improves the efficiency and accuracy of determining the actual position information of the target object in the image.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of image processing, and in particular, to a method, apparatus, device, and storage medium for identifying a motion trajectory of an object. Background technique [0002] With the rapid development of image processing technology, more and more fields use images to monitor, process and analyze objects. How to accurately determine the location information of objects in the image is particularly important. [0003] In the prior art, images are collected through image collection devices such as cameras, and processing methods such as image recognition are used to identify targets and locate objects in images, or combine manual processing to determine objects such as buildings or roads in images, and use video images to identify objects such as buildings or roads. Layer and mark technology, superimpose mark information and camera point information on the video to judge the corresponding rel...

AI Technical Summary

Problems solved by technology

[0004] However, the determination of the target position in the prior art is still limited to the range of the image, and the image position of the target can only be converted into the geographic location in the actual scene by means of manual means, so the efficiency and accuracy of determining the actual position of the target in the image are relatively low. Low

Images