【摘要】：人臉肖像繪制機(jī)器人是一種可以自動(dòng)繪制人臉肖像輪廓圖的智能機(jī)械系統(tǒng),它由圖像采集模塊、圖像處理模塊、機(jī)械控制繪圖模塊組成,能夠自動(dòng)拍攝人臉照片,提取肖像輪廓,并控制機(jī)械臂在畫板上畫出人臉線條畫。人臉肖像繪制機(jī)器人是機(jī)器視覺的研究方向之一,可廣泛用于科普展覽,其中提出的基于機(jī)器視覺的研究技術(shù)在生產(chǎn)和生活等各個(gè)方面都有著廣泛的應(yīng)用。本文主要的研究內(nèi)容和成果有: 研究了圖像采集系統(tǒng)的硬件和軟件算法,通過對(duì)CCD攝像機(jī)、光學(xué)鏡頭、圖像采集卡的選型和拍攝算法的設(shè)計(jì),完成自動(dòng)采集和保存人臉肖像照片的功能。在對(duì)基于膚色的圖像分割技術(shù)研究的基礎(chǔ)上,對(duì)比目前常用的色彩空間的原理和特點(diǎn),采用YCbCr色彩空間進(jìn)行圖像分解,并提出了與迭代閾值法相結(jié)合的圖像分割方法進(jìn)行膚色分割,研究了圖像輪廓提取的一般方法,并使用內(nèi)部掏空法提取出人臉輪廓。 研究了人臉五官特征提取的一般方法,提出一種基于搜索的投影法對(duì)眼睛和眉毛進(jìn)行坐標(biāo)定位,研究去除噪聲提取眼睛和眉毛的搜索算法,優(yōu)化處理人臉輪廓圖,對(duì)改善人臉肖像圖的整體效果達(dá)到較好效果。 提出一種基于無干擾八領(lǐng)域搜索的矢量化算法,將人臉輪廓像素點(diǎn)轉(zhuǎn)化為一系列相連的點(diǎn),再檢測(cè)每組相連的點(diǎn)中位于一條直線上的點(diǎn),保存線段首位點(diǎn)坐標(biāo)到數(shù)組中,為與機(jī)械臂控制器通信做好準(zhǔn)備。 研究繪圖機(jī)械控制系統(tǒng)的硬件選型和控制算法,解決了外部坐標(biāo)值與控制器內(nèi)部變量通信的技術(shù)難點(diǎn),在Visual C++ 6.0中實(shí)現(xiàn)了外部對(duì)機(jī)械臂繪圖動(dòng)作的自動(dòng)控制,設(shè)計(jì)機(jī)械臂繪畫動(dòng)作流程,完成人臉輪廓圖的自動(dòng)繪制。
[Abstract]:The face portrait drawing robot is an intelligent mechanical system which can automatically draw the profile of the face portrait. It is composed of image acquisition module, image processing module and mechanical control drawing module, which can automatically take face photos. Extract the profile of the portrait and control the robotic arm to draw the face line on the drawing board. Face portrait rendering robot is one of the research directions of machine vision, which can be widely used in popular science exhibition. The research technology based on machine vision has been widely used in production and daily life. The main contents and achievements of this paper are as follows: the hardware and software algorithms of the image acquisition system are studied, and the selection of the CCD camera, optical lens, image acquisition card and the design of the shooting algorithm are discussed. The function of automatic collection and preservation of face portrait photos is completed. Based on the research of image segmentation technology based on skin color, this paper compares the principle and characteristics of current color space, and uses YCbCr color space to decompose the image. An image segmentation method combined with iterative threshold method is proposed, and the general method of image contour extraction is studied, and the face contour is extracted by internal tunneling method. In this paper, the general method of facial feature extraction is studied, and a search-based projection method is proposed to locate the coordinates of the eyes and eyebrows. The search algorithm for removing noise from the eyes and eyebrows is studied, and the face contour image is optimized. It can improve the overall effect of face portrait. In this paper, a vectorization algorithm based on non-interference eight domain search is proposed, which transforms the face contour pixels into a series of connected points, then detects the points on a straight line in each group of connected points, and saves the coordinates of the first point of the line segment into the array. Prepare for communication with manipulator controller. This paper studies the hardware selection and control algorithm of drawing mechanical control system, solves the technical difficulty of communication between external coordinate value and internal variable of controller, and realizes the automatic control of drawing action of mechanical arm in Visual C 6.0. The drawing process of the robot arm is designed, and the automatic drawing of the face contour is completed.
【學(xué)位授予單位】：中國科學(xué)技術(shù)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2011
【分類號(hào)】：TP242

【參考文獻(xiàn)】

相關(guān)碩士學(xué)位論文 前2條

1 周海波;基于活動(dòng)輪廓模型的人臉輪廓提取方法的研究[D];四川大學(xué);2004年

2 程勇光;基于膚色和面部幾何特征的人臉檢測(cè)算法的研究[D];北京交通大學(xué);2008年

，

本文編號(hào)：2376900