A Contour Detection Method for Indoor Scenes by Fusion of Color and Depth Information

Abstract

The invention discloses an indoor scene outline detection method fusing color and depth information and belongs to the technical field of machine vision. The method comprises the key techniques of 1,separating a color image into R (red), G (green) and B (blue), using an image edge detection operator to extract edge information on each channel, and performing weighted fusing to obtain overall results of color image edge detection; 2, using a depth image to extract edge information, correcting false edges due to the loss of three-dimensional information of a scene only during detection of the color image, and enhancing confidence of true edge detection results; 3, constructing a gradient pyramid, and performing multi-scale multi-level fusing on color image edges and depth image edges to obtain edge detection results, and executing edge aggregation to obtain a final outline. The method of the invention has the advantages that gradient information of a depth image is fully excavated, outline detection results of a color image are corrected and enhanced via the outline detection results of the depth image, and a more robust and precise scene outline is obtained accordingly.

Description

technical field [0001] The invention belongs to the technical field of machine vision, and more specifically relates to an indoor scene contour detection method that combines color and depth information. Background technique [0002] In recent years, technologies or applications such as intelligent robots, virtual reality, and drones have become more mature and are rapidly integrating into people's daily production and life. These applications rely on various technologies in the field of machine vision, including object detection and recognition, scene classification, etc., and the more basic one is to extract image edge information to form contours, which is one of the prerequisites for smart devices to understand scenes. How to obtain more accurate and human-perceived object outlines from two-dimensional images acquired by sensing devices is currently a research hotspot in academia and industry. Among them, indoor robots, which are more closely related to human life, are ...

AI Technical Summary

Problems solved by technology

For each pixel, a posterior probability between 0 and 1 is obtained, and then the watershed transform algorithm is used to convert these inputs into closed areas, and there is a problem in this way, that is, there are strong boundaries near some weak boundaries, which are affected by them. The boundary is also misjudged as a contour, resulting in an error in the final contour detection result

Images