Real-time object identification method suitable for indoor environment

Abstract

The invention discloses a real-time object identification method suitable for indoor environment. The method comprises the following contents of an offline training stage and an online identificationstage, wherein the offline training stage comprises the following steps that: obtaining the point cloud data of a target object to be identified, calculating the geometrical characteristic descriptionof the point cloud data of the target object to be identified, obtaining a characteristic histogram, and constructing an offline model database according to the characteristic histogram. The online identification stage comprises the following steps that: collecting data, and carrying out point cloud filtering, clustering segmentation and matching identification. By use of the method, a normal angle relationship between the center point of the point cloud data of the object and all other points on the surface is used for showing the geometrical characteristic of an object surface, other pointsin the neighbourhood of the point is coded according to the angle deviation between each point and the normal directions of other points in the neighbourhood of the point, and the weight of each point in the surface geometrical characteristic description of the point cloud data of the object is calculated to obtain the geometrical shape characteristic histogram of the object surface. The method is not influenced by ray changes, and object identification accuracy is improved.

Description

technical field [0001] The invention relates to a real-time object recognition method suitable for indoor environment. technical background [0002] With the continuous advancement of science and technology, artificial intelligence is becoming more and more popular, and technologies such as computer vision, robotics, and augmented reality continue to develop, and object recognition technology is also particularly important. Traditional object recognition technology relies on two-dimensional visual images, and its recognition effect is easily affected by lighting conditions, and its application is even more difficult when the brightness of the light changes or at night. The object recognition technology based on 3D point cloud can not be affected by the lighting conditions. The 3D point cloud provides the depth information in the scene. It can perform scene recognition more effectively, and then accurately identify targets. But in some real-time applications, its recognitio...

AI Technical Summary

Problems solved by technology

However, in some real-time applications, its recognition accuracy and real-time performance are difficult to meet the requirements of practical applications.

Images