Font Size: a A A

Research On Gesture Control Of Security Robot

Posted on:2021-05-20Degree:MasterType:Thesis
Country:ChinaCandidate:Y F WuFull Text:PDF
GTID:2518306512979009Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Due to the recent enthusiasm of new concepts and technologies such as Cyber-Physical System(CPS)and Internet of Things(IO T),human-computer interaction technology has become increasingly important.These new technologies revolve around the theme of combining the real world and the information world into one.In the process,gestures have gained a lot of attention as a human-computer interaction method that can be conveniently used by the public with almost no learning.This article studies the gesture control of security robots.The main research contents and innovations are as follows:(1)Research on the security robot control system,analyze the characteristics and control requirements of the security robot,and design a set of control gestures that are easy for users to familiarize with,easy to operate,and low misoperation.To solve the problem that gestures have different meanings in different directions,a gesture direction recognition algorithm is designed to prioritize the gesture direction before gesture recognition to solve this problem.(2)The gesture segmentation algorithm based on skin color model is researched and improved.The skin color segmentation and inter-frame contrast are merged to reduce the noise of the skin color segmentation algorithm and improve the success rate of gesture segmentation.In the case of occlusion,a morphological operation that can fill t he occluded part when the occlusion size is unknown is proposed.(3)Research and improve the current finger detection algorithm,apply convex hull points and convex defect points to fingertip detection to extract the number of fingers.(4)Analyze the geo metric and statistical characteristics such as Hu moment,density,rectangularity,elongation,roundness,perimeter ratio,number of fingers,etc.using gray theory,test different characteristics and use gray theory to analyze,Find the best combination of features.(5)For the multi-classification problem of support vector machines,combined with the actual situation of security robot gesture control,based on the indirect method of one-to-one classification,an inseparable situation was introduced,and t he test samples were judged to be invalid data by screening the votes.
Keywords/Search Tags:gesture set design, gesture segmentation, fingertip detection, feature analyze
PDF Full Text Request
Related items