本文選題：機(jī)器人視覺 + 特征點提取　； 參考：《昆明理工大學(xué)》2017年碩士論文

【摘要】：視覺的引入是機(jī)器人智能化的一個巨大飛躍,機(jī)器人視覺標(biāo)定技術(shù)是實現(xiàn)機(jī)器人視覺伺服控制、引導(dǎo)機(jī)器人運動和測量的前提,由此可見機(jī)器人視覺標(biāo)定對機(jī)器人完成相關(guān)動作至關(guān)重要。本文主要就攝像機(jī)視覺標(biāo)定技術(shù)在機(jī)器人視覺引導(dǎo)方面進(jìn)行研究,機(jī)器人視覺標(biāo)定精度直接影響到視覺引導(dǎo)的準(zhǔn)確性,而圖像特征點檢查精度又直接影響攝像機(jī)標(biāo)定精度,因此重點研究了攝像機(jī)標(biāo)定過程中圖像特征點檢查技術(shù),本文的主要研究工作如下:1、從應(yīng)用角度出發(fā),提出了一種改進(jìn)的Harris方格板角點檢測算法。該算法綜合運用了 Harris角點檢測技術(shù),并在保留了 Harris角點檢測算法良好的可重復(fù)性和相對較高的檢測效率的情況下使其精度和可重復(fù)性更高,同時很好的解決了相機(jī)內(nèi)參數(shù)計算過程中:方格板圖像的角點坐標(biāo)和空間點相匹配這一難點問題,并有效的解決了原有Harris角點檢測算法閾值的選取過度依賴的問題。2、提出了一種基于幾何對稱性并應(yīng)用于橢圓陣列圖像的圓心檢測算法,該算法通過設(shè)定約束條件來限定搜索范圍,避免出現(xiàn)漏檢或相互混淆而影響檢測的精度;同時有效的解決了標(biāo)定板的標(biāo)記點和圖像特征點難以嚴(yán)格匹配的問題。3、尋求一種完全自動的圓心特征提取算法,達(dá)到降低機(jī)器人標(biāo)定對工人的文化水平需求,從而降低企業(yè)成本的目的;并針對現(xiàn)有相關(guān)技術(shù)中同時對多個橢圓進(jìn)行檢測時出現(xiàn)精度不高和難以準(zhǔn)確排序的問題,提出一種將最小二乘法應(yīng)于橢圓陣列圖像的橢圓擬合算法,該算法無需人機(jī)交互協(xié)同完成檢測工作,全部自動完成提取,提取到的圓心特征點精度和匹配度相對較高,可重復(fù)性好。4、利用上述的三種特征點提取算法利用平面標(biāo)定法進(jìn)行實驗,計算出攝像機(jī)的內(nèi)外參數(shù),并進(jìn)行對比分析;將標(biāo)定數(shù)據(jù)寫入機(jī)器人控制器,進(jìn)行視覺抓取實驗進(jìn)一步驗證標(biāo)定算法的正確性。
[Abstract]:The introduction of vision is a great leap of robot intelligence. Robot vision calibration technology is the premise to realize robot vision servo control, guide robot motion and measurement. It can be seen that robot vision calibration is very important for robot to complete related actions. In this paper, the camera vision calibration technology in robot vision guidance is studied. Robot vision calibration accuracy directly affects the accuracy of vision guidance, and image feature point inspection accuracy directly affects camera calibration accuracy. Therefore, this paper focuses on the image feature point detection in camera calibration. The main work of this paper is as follows: 1. From the point of view of application, an improved Harris grid corner detection algorithm is proposed. In this algorithm, the Harris corner detection technique is used synthetically, and the accuracy and repeatability of the Harris corner detection algorithm are improved by keeping the good repeatability and relatively high detection efficiency of the Harris corner detection algorithm. At the same time, it solves the difficult problem of matching corner coordinates and space points in the calculation process of camera inner parameters. The problem of over-dependence of threshold selection of the original Harris corner detection algorithm is effectively solved, and a circle center detection algorithm based on geometric symmetry and applied to elliptical array images is proposed. The algorithm limits the search range by setting the constraint conditions to avoid the error detection or confusion and affects the accuracy of the detection. At the same time, it effectively solves the problem that the mark points and image feature points of the calibration board are difficult to match strictly, and seeks a completely automatic center feature extraction algorithm to reduce the requirement of robot calibration for workers' cultural level. In order to reduce the cost of the enterprise, aiming at the problems of low precision and difficult to sort the ellipses in the existing related technology, an ellipse fitting algorithm is proposed, which applies the least square method to the elliptic array image. This algorithm does not need man-machine interaction cooperation to complete the detection work, all of which are automatically extracted, and the accuracy and matching degree of the extracted centroid feature points are relatively high. The method of plane calibration is used to carry out experiments, to calculate the internal and external parameters of the camera, to compare and analyze, to write the calibration data into the robot controller, and to write the calibration data into the robot controller. The calibration algorithm is verified by visual capture experiment.
【學(xué)位授予單位】：昆明理工大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前6條

1 林明盛;吳磊;柯科勇;;視覺測量中攝像機(jī)標(biāo)定精度評估方法[J];機(jī)床與液壓;2016年23期

2 陳愛珍;;日本工業(yè)機(jī)器人的發(fā)展歷史及現(xiàn)狀[J];機(jī)械工程師;2008年07期

3 鄒益民;汪渤;;一種基于最小二乘的不完整橢圓擬合算法[J];儀器儀表學(xué)報;2006年07期

4 劉振宇,陳英林,曲道奎,徐方;機(jī)器人標(biāo)定技術(shù)研究[J];機(jī)器人;2002年05期

5 吳福朝,李華,胡占義;基于主動視覺系統(tǒng)的攝像機(jī)自定標(biāo)方法研究[J];自動化學(xué)報;2001年06期

6 李華,吳福朝,胡占義;一種新的線性攝像機(jī)自標(biāo)定方法[J];計算機(jī)學(xué)報;2000年11期

相關(guān)碩士學(xué)位論文 前4條

1 曹焱;基于雙目立體視覺信息的三維重建方法研究[D];吉林大學(xué);2013年

2 范東凱;攝像機(jī)自標(biāo)定的算法研究[D];哈爾濱工業(yè)大學(xué);2010年

3 陳鐸;攝像機(jī)標(biāo)定方法研究與實現(xiàn)[D];東北大學(xué);2008年

4 陳靜;圖像配準(zhǔn)特征點提取算法研究[D];南京理工大學(xué);2006年

，

本文編號：1860303