Scene modeling method under single camera and system

Abstract

The invention relates to a scene modeling method under a single camera and a system. The method comprises the following steps of: detecting motion targets according to a video frame image; tracking every detected motion target to obtain motion trails of the motion targets; screening out calibration object targets from the motion trails of all the motion targets; selecting at least three positionsfrom the screened motion trails of the motion targets as calibration point positions; extracting image coordinates of highest points and lowest points of the imaging of the calibration object targetsat the calibration point positions; calculating the positions of vertical vanishing points and horizontal vanishing lines under a current scene through the coordinates and the mounting height of the camera; and building a height model. The invention does not need artificial auxiliary calibration and can automatically carry out modelling, be self-adaptive to the change of the scene and simultaneously process multiple videos of different scenes.

Description

technical field [0001] The invention relates to the field of scene modeling, in particular to a method and system for scene modeling under a single camera. Background technique [0002] In order to obtain 3D spatial information from 2D images, scene modeling is an essential step, and it is widely used in 3D reconstruction, navigation, visual monitoring and other fields. Reconstructing the true size of an object from an unmarked image sequence is a very meaningful technique with broad application prospects. For example, in the automatic navigation task of the robot, the robot obtains the two-dimensional image of the environment through observation, and then obtains the three-dimensional information of the environment through calculation to determine the walking direction, distance, height of the target in the field of view, etc.; Through scene modeling, information such as the size and speed of moving objects in the scene can be calculated, which is convenient for timely mon...

AI Technical Summary

Problems solved by technology

[0004] (1) It is necessary for a person to enter the scene with a calibration board, which may be inconvenient or impossible in some application environments

[0005] (2) It is necessary to manually select the imaging points of all the points on the calibration board in each orientation to determine the image coordinates of these points, which is cumbersome and error-prone

[0006] (3) The selected image coordinates cannot be optimal

Especially when the calibration board is small, the dot matrix imaging on the board is less obvious, which will seriously affect the positioning of the image coordinates

[0007] (4) Cannot adapt to scene changes

Once the model is successfully established, it is required that the camera cannot be changed. When the camera changes and the shooting scene changes (such as: the position of the camera shifts, or the focal length changes, etc.), the model must be re-established, otherwise it will seriously affect the calculation. result

[0008] (5) Multiple videos of different scenes cannot be processed at the same time

[0009] From the above five points, it can be seen that the current scene modeling technology is not flexible, and even cannot be implemented in some occasions, which brings great difficulties to video applications such as surveillance

Images