本文選題：交通錐 + 圖像處理��； 參考：《鄭州大學(xué)》2017年碩士論文

【摘要】：本文從數(shù)字圖像處理角度出發(fā),對(duì)交通錐的識(shí)別與定位算法進(jìn)行了深入研究。該算法主要用于輔助交通錐收放車對(duì)交通錐進(jìn)行自動(dòng)識(shí)別與定位,有利于提高交通錐收放車的收錐效率和自動(dòng)化性能。論文主要內(nèi)容包括以下三個(gè)方面:(1)依據(jù)交通錐的顏色和形狀特性,提出了一種基于連通域大小和特征點(diǎn)位置關(guān)系的識(shí)別算法。首先基于HSV顏色空間對(duì)獲取的交通錐圖像進(jìn)行分割,然后利用形態(tài)學(xué)方法對(duì)顏色分割后的二值圖像進(jìn)行一系列處理,包括基于連通域大小的目標(biāo)篩選、輪廓提取和基于輪廓外接矩形特征點(diǎn)的分析,以實(shí)現(xiàn)交通錐的識(shí)別。(2)基于幾何關(guān)系推導(dǎo)法,對(duì)交通錐的定位進(jìn)行了建模。依據(jù)交通錐紅色區(qū)域的特征點(diǎn),根據(jù)攝像機(jī)的投影模型,結(jié)合它們之間的幾何關(guān)系,推導(dǎo)出交通錐在路面上的位置信息。(3)根據(jù)交通錐的識(shí)別與定位結(jié)果,為后續(xù)機(jī)械臂的抓取建立交通錐角度的模板庫�；谔卣鼽c(diǎn)的位置關(guān)系進(jìn)行交通錐倒下狀態(tài)時(shí)角度的測(cè)算,由于交通錐圖片是帶有高度的攝像頭拍攝到的,三維信息在轉(zhuǎn)化成二維圖片中交通錐角度的過程中由于缺少深度信息從而造成誤差。因此需要建立一個(gè)映射的模板庫,模板庫中包含了從二維圖像中的角度到機(jī)械臂抓取角度的映射。為了對(duì)本文算法進(jìn)行驗(yàn)證,在Visual Studio 2010+OpenCV平臺(tái)上開發(fā)了相應(yīng)的應(yīng)用程序,對(duì)以上的研究內(nèi)容進(jìn)行了較為系統(tǒng)的實(shí)驗(yàn)測(cè)試。實(shí)驗(yàn)結(jié)果表明,該算法能夠有效辨認(rèn)出路面上的交通錐,實(shí)驗(yàn)中交通錐的識(shí)別率為97.8%,交通錐的定位誤差均小于等于2厘米,交通錐角度的測(cè)量誤差小于16°,能夠基本滿足交通錐收放車上機(jī)械臂對(duì)交通錐抓取工作的精度要求。
[Abstract]:From the angle of digital image processing, this paper studies the recognition and location algorithm of traffic cone in depth. This algorithm is mainly used to assist the automatic recognition and positioning of traffic cones for traffic cones. It is beneficial to improve the efficiency and automatic performance of the cones. The main contents of this paper include the following three aspects: (1) According to the color and shape characteristics of the traffic cones, a recognition algorithm based on the relationship between the size of the connected domain and the location of the feature points is proposed. First, a traffic cone image is segmented based on the HSV color space, and then a series of two value images are processed by the morphological method, including the target based on the size of the connected domain. The identification of traffic cones is realized by screening, contour extraction and rectangle feature points based on contour lines. (2) based on geometric relation derivation, the positioning of traffic cones is modeled. According to the feature points of the traffic cone red area, the traffic cone is derived on the road according to the projection model of the camera and the geometric relationship between them. Position information. (3) according to the recognition and positioning results of the traffic cone, the template library of the traffic cone angle is set up for the capture of the following manipulator. Based on the position relation of the feature points, the angle of the traffic cone falls down, because the traffic cone picture is photographed with the height of the camera, and the three-dimensional information is converted into the two dimensional picture. In the process of cone angle, the error is caused by lack of depth information. Therefore, a mapping template library is needed. The template library contains the mapping from the angle of the two-dimensional image to the grasping angle of the manipulator. In order to verify this algorithm, the corresponding application is developed on the Visual Studio 2010+OpenCV platform. The experimental results show that the algorithm can effectively identify the traffic cones on the road, the recognition rate of the traffic cone in the experiment is 97.8%, the positioning error of the traffic cone is less than 2 cm, and the measurement error of the angle of the traffic cone is less than 16 degrees, and it can basically meet the machinery on the traffic conical retractable car. The precision requirements of the arm to catch the traffic cone.
【學(xué)位授予單位】：鄭州大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 趙書俊;段紹麗;張曉芳;李磊;劉曉e，

本文編號(hào)：1927380