Heterogeneous depth camera-based three-dimensional information obtaining method and apparatus

Abstract

The invention discloses a heterogeneous depth camera-based three-dimensional information obtaining method. The method comprises the following steps of shooting a picture; calculating internal parameters of a Kinect camera and a TOF camera; extracting reliable depth information according to the shot picture; fitting calibration board plane coordinates of the Kinect camera and the TOF camera according to the reliable depth information; recovering an actual three-dimensional coordinate set of corner points of the Kinect camera and the TOF camera under the fitted calibration board plane coordinates in combination with the internal parameters; calculating a least square conversion matrix between the Kinect camera and the TOF camera; performing coordinate conversion on the reliable depth information according to the least square conversion matrix to obtain depth fusion information; matching color information acquired by the Kinect camera with the depth fusion information; and performing point cloud registration on image information after matching based on a spatial distance between the points to obtain three-dimensional information. By adopting the method, a high-resolution accurate depth map fusing the color information can be quickly obtained, and then accurate three-dimensional information is obtained.

Description

technical field [0001] The invention belongs to the field of image processing and point cloud registration, and in particular relates to a method and device for acquiring three-dimensional information based on heterogeneous depth cameras. Background technique [0002] At present, there are two main ways to obtain the depth information of objects: the stereo matching method and the depth camera method to obtain the depth by calculating the disparity of image matching points. Among them, the widely used depth cameras mainly include Kinect cameras using structured light technology (referring to Microsoft’s Kinect XBOX360, the same below) and TOF cameras using Time of Flight (Time of Flight) technology (represented by Mesa’s SR4000). Both depth cameras have their own advantages and disadvantages: the TOF camera can generate a depth map in real time, and the depth information has high precision and can support hardware triggering, but the image resolution is low. For example, th...

AI Technical Summary

Problems solved by technology

Both depth cameras have their own advantages and disadvantages: the TOF camera can generate a depth map in real time, and the depth information has high precision, and can support hardware triggering, but the image resolution is low

Its shortcoming is that ordinary color cameras are greatly affected by illumination and texture, so the operability of direct calibration between color cameras and TOF depth cameras is poor

However, the accuracy of similar control sets found only by geometric features is still difficult to guarantee, and the number of control sets is also very large

Images