Depth map optimization method and device based on semantic instance segmentation, equipment and medium

Abstract

The invention provides a depth map optimization method and device based on semantic instance segmentation, equipment and a medium. The method comprises the following steps: acquiring a light field image set, obtaining a scene depth map by using a light field depth calculation method, detecting one or more objects in each image by using a semantic instance segmentation matching algorithm for the images acquired by each camera, and matching the one or more objects to corresponding pixel regions and position frames; calculating the object center of the corresponding object in the corresponding camera according to the position frame, and obtaining an approximate depth value according to the relative position relationship between the cameras; and calculating an object depth map corresponding tothe object according to the approximate depth value, and fusing the object depth map into the scene depth map to obtain a final depth map corresponding to each object. According to the invention, rapid calculation of the depth map can be realized in a complex scene environment, a more accurate depth map can be obtained under the condition of the same calculation amount, and meanwhile, the balancebetween the calculation speed and the precision can be adjusted.

Description

technical field [0001] The present invention relates to the technical field of computer vision processing, in particular to a depth map optimization method, device, equipment and medium based on semantic instance segmentation. Background technique [0002] With the popularity of machine vision, automatic driving, and robots, it is particularly important to obtain the depth of the scene; in some applications, such as object recognition and obstacle detection in automatic driving, recognition, sorting, Depalletizing, palletizing, and shelf grabbing of objects in logistics scenarios, etc., all require the use of depth map information. There are two main ways to obtain the depth map. One is to obtain the depth map through the depth camera, but it is often limited by the accuracy, effective distance and cost of the depth camera. The other is binocular stereo vision, which is an important form of machine vision. It is based on the principle of parallax and uses imaging equipment ...

AI Technical Summary

Problems solved by technology

However, the traditional binocular stereo vision will affect the result of scene depth due to changes in scene lighting, which will reduce the accuracy of the depth map or generate noise.

Images