Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Real-time object detection method based on combination of three-dimensional point cloud segmentation and local feature matching

A three-dimensional point cloud and local feature technology, applied in computer parts, character and pattern recognition, instruments, etc., can solve problems such as unsatisfactory results, and achieve the effect of good image recognition, simple calculation, and fast processing speed.

Inactive Publication Date: 2017-03-22
BEIJING UNION UNIVERSITY
View PDF3 Cites 27 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method is usually affected by factors such as the limited dynamic range of the camera itself and changes in the external environment, and the effect of object detection is not ideal.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Real-time object detection method based on combination of three-dimensional point cloud segmentation and local feature matching
  • Real-time object detection method based on combination of three-dimensional point cloud segmentation and local feature matching
  • Real-time object detection method based on combination of three-dimensional point cloud segmentation and local feature matching

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0047] In view of the shortcomings of previous methods and the needs of service robots in practical applications, this paper proposes a real-time object recognition method that combines 3D point cloud and color image processing, through the rapid segmentation of 3D point cloud to achieve object detection and positioning, and uses The method based on SURF feature matching realizes the recognition of the target object. Taking advantage of the complete structural characteristics of ordered point cloud data, a method combining integral images and region growing is proposed to achieve faster 3D object detection. This method performs special matching in two steps, and the input is a 3D point cloud and a color image, which is divided into two steps. The first step is to realize object detection. Since robot-operable objects (such as drinks, cups, pen holders, etc.) in indoor environments are usually placed on horizontal surfaces such as tables and shelves, we first segment these hori...

Embodiment 2

[0063] Such as figure 2 , 2a , 2b, 2c, 2d, 2e, figure 2 There are three barrel-shaped objects with different designs and colors in the picture, one is a red coke can with patterns, one is a blue potato chip barrel with patterns, and the other is a water glass with a handle and a white bottom with curved patterns. The first step is to carry out horizontal clustering, and the figure 2 The RGB-D original image shown in is segmented horizontally to obtain Figure 2a , first make sure that the three barrel-shaped objects are on the same horizontal plane I. The second step is to perform object clustering, and the Figure 2a The image shown in the object segmentation, get Figure 2b , remove the scenery around the three barrel-shaped objects, keep only the three barrel-shaped objects, extract the background color, pattern and whether there are protrusions on each barrel-shaped object, that is, the characteristics of the object, and store it in the feature library . Step thr...

Embodiment 3

[0065] Such as image 3 As shown in , in the 3D point cloud of the image, I represents a segmented horizontal plane, II represents the boundary of the horizontal plane, and III represents the point cloud of three objects on the horizontal plane.

[0066] Such as Figure 3a shown, given image 3 The point cloud segmentation results of three objects at III on the horizontal plane at I. Among them, IV represents three separate object clusters that are separated from each other.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a real-time object detection method based on combination of three-dimensional point cloud segmentation and local feature matching. The method comprises the following steps that: a normal vector of each point in a three-dimensional point cloud is calculated rapidly by combining an integral image so as to carry out horizontal plane segmentation on an object; the object is segmented; target object feature extraction is carried out; and the extracted target object feature is compared with an object feature database. According to the method provided by the invention, object detection and positioning are realized by rapid segmentation of the three-dimensional point cloud; and with an SURF-feature-matching-based method, target object identification is realized. Because of the integrated structure characteristics of orderly point clod data, a method combining the integral image with regional growth is provided, so that rapid three-dimensional object detection is realized.

Description

technical field [0001] The invention relates to the technical field of digital image processing, in particular to a real-time object detection method combining three-dimensional point cloud segmentation and local feature matching. Background technique [0002] The research on object classification and detection is the basis of the entire computer vision research and the basis for solving other complex visual problems such as tracking, segmentation, and scene understanding. Therefore, if we want to automatically analyze and understand actual complex scenes, we first need to determine what objects exist in the image (classification problem), or determine what objects exist in what position in the image (detection problem). [0003] Prior to this, SwainMJ, BallardDH. proposed using the color histogram method to describe and recognize objects in "Colorindexing", but this method is greatly affected by illumination; later LoweDG. The SIFT feature proposed in "Object recognition fr...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06K9/46
CPCG06V10/443G06F18/22
Inventor 袁家政刘宏哲郭燕飞
Owner BEIJING UNION UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products