A Method of Object Detection in Depth Image

Abstract

The invention relates to a method for detecting objects in a depth image, comprising the following steps: establishing a depth tree model based on a multi-fork tree data structure and a mapping diagram between depth tree nodes and depth image pixels for an input depth image; traversing each node of the depth tree model, Statistics and analysis of node data, judgment of node attributes, so as to obtain the rectangular area where the object is located. The invention can quickly and accurately obtain the rectangular area of ​​the object, and improve the accuracy of object detection in the depth image, so that the detection result can not only meet the visual requirements of the human eye, but also better meet the requirements for speed and precision in practical applications.

Description

technical field [0001] The invention relates to object detection technology in the field of computer vision, in particular to a method for object detection in depth images. Background technique [0002] Object detection has always been a basic problem in the field of computer vision, and it is also a difficult problem. Obtaining as few object regions as possible and independent of object categories is the core of object detection. Traditional methods are mostly based on rectangular frames, that is, a series of rectangular frames are used to frame the objects in the image as tightly as possible. This kind of method is widely used in the traditional two-dimensional image field, and has achieved good detection accuracy while taking into account the speed. . Another method is based on region segmentation, that is, segmenting the object region with specific practical significance in the image. The accuracy of this method is generally high, but the algorithm complexity is high a...

AI Technical Summary

Problems solved by technology

This type of method is widely used in the field of traditional two-dimensional images, and has achieved good detection accuracy while taking into account the speed. However, this method also has problems such as generating more candidate boxes and low compactness of the candidate boxes.

Another method is based on region segmentation, which is to segment out object regions with specific practical significance in the image. This type of method generally has high accuracy, but the algorithm complexity is high and the speed is generally slow.

Images