Unlock instant, AI-driven research and patent intelligence for your innovation.

A multi-feature combined robot football recognition method

A robot, multi-feature technology, applied in character and pattern recognition, instruments, computer parts, etc., can solve the problems of complex stadium background, misjudgment, large stadium range, etc., to achieve high accuracy and robust performance and real-time performance , responsive and less time-consuming

Active Publication Date: 2019-07-26
ZHEJIANG UNIV OF TECH
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The recognition effect of the traditional method is greatly reduced; 2. The background of the stadium is complex, the white line of the stadium, the goal, and the crowd cause great interference to the recognition method. The method often detects the complex background while detecting the ball, resulting in misjudgment; 3. The range of the court is large, and the scales of short-range balls and long-distance balls are quite different, or even doubled in size. It is difficult to achieve consistent recognition by traditional methods

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
  • A multi-feature combined robot football recognition method
  • A multi-feature combined robot football recognition method
  • A multi-feature combined robot football recognition method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0021] The present invention will be further described below in conjunction with the accompanying drawings.

[0022] A multi-feature joint robot football recognition method, comprising the steps of:

[0023] Step 1. Preprocessing. Such as image 3 , use the humanoid robot camera to collect stadium pictures in different environments (illumination, background, viewing angle), manually calibrate, and obtain positive samples (soccer pictures), including 64x64 and 32x32 sizes; the rest collect enough negative samples (non- Football pictures), the number is roughly 3-5 times that of the positive samples, including two sizes of 64x64 and 32x32.

[0024] Step 2. Training. Such as image 3 , for the above 64x64 positive samples and negative samples, after extracting HoG features, train 64x64 window SVM_HoG_64, extract RGBSIFT features and use BoW to get a unified feature descriptor, and then train 64x64 window SVM_RGBSIFT_64, the size classifier is used to detect the position of th...

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

A multi-feature joint football position recognition method, based on feature extraction and classification, improves the traditional one-time detection method, and adopts a long-short distance multi-scale secondary detection strategy. Use the common HoG+SVM algorithm for target detection as the initial detection model, extract a series of preliminary calibration frames (may contain background interference), and use RGB‑SIFT+ Bag of Words (Bag of Words)+SVM for the second screening for secondary detection Obtain the target calibration frame, and finally obtain the final calibration frame through non-maximum suppression. This method has high recognition accuracy and sensitive response for humanoid robot.

Description

technical field [0001] The present invention relates to a real-time recognition method of football in a RoboCup humanoid robot complex stadium environment. Specifically, a humanoid robot camera is used to take scene pictures, and multiple types of features (RGB-SIFT, HoG) are extracted through a sliding window, and a bag of words model (Bag of Words) to form a unified feature descriptor, which is classified by a support vector machine (SVM), and finally determines the actual position of the football and feeds it back to the robot. This method belongs to the field of computer vision / object detection. Background technique [0002] Robot football is a test of artificial intelligence in a dynamic and uncertain environment. It is a high-tech confrontation based on sports competitions. It is an important means of cultivating scientific and technological talents in the field of information and automation. An effective way to achieve practical and industrialization. The research f...

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 Patents(China)
IPC IPC(8): G06K9/62G06K9/66
CPCG06V30/194G06F18/2411
Inventor 赵燕伟朱炎亮屠海龙盛尚浩鞠振宇赵晓
Owner ZHEJIANG UNIV OF TECH