Depth image generating method and apparatus, and electronic device

Abstract

The embodiment of the invention provides a depth image generating method and apparatus, and an electronic device. The method comprises: a first color image and a first depth image that include same scenery are obtained; according to similarity between reference pixel points and non-reference pixel points respectively included in a plurality of first preset regions in the first color image, the pixel points included by the plurality of first preset regions are classified; according to pixel points being located in the same first preset regions and having the same types, corresponding target pixel points of the reference pixel points included by the plurality of first preset regions in the first depth image are determined; the depth values of the target pixel points are determined as depth coefficients of to-be-generated pixel points in a second depth image; and a second depth image is generated by combining the depth coefficients. Therefore, the target pixel points are determined basedon the pixel point types and thus the depth values of the to-be-generated pixel points are calculated accurately, so that the accuracy of the generated depth image is improved and an object distortionphenomenon occurring in the virtual viewpoint image is avoided.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a depth image generation method, device and electronic equipment. Background technique [0002] There are usually some special working scenarios or working environments in different fields, such as computer-assisted surgery in the medical field or working environments that are likely to cause harm to the human body in the industrial field. In the above cases, people usually need to use robots to complete the preset work. The robot can successfully complete the preset work based on a comprehensive understanding of the working environment, but the limited number of camera devices configured on the robot cannot achieve a comprehensive understanding of the working environment, thereby reducing the success of the preset task. Rate. [0003] In order to overcome this defect, in practical applications, a virtual viewpoint synthesis algorithm is usually used to synthes...

AI Technical Summary

Problems solved by technology

However, the accuracy of the high-resolution depth image obtained by using the above algorithm is usually low, and it cannot accurately represent the positional relationship between objects, which leads to the phenomenon of object distortion in the virtual viewpoint image generated by using the depth image.

Images