目 录
摘 要 I
ABSTRACT II
1 引言 1
1.1 研究背景 1
1.2 机器人发展史 1
1.3 机器人视频导航概述 2
1.4 世界各国机器人研究与应用 2
1.5 机器人导航未来发展方向 3
1.6 本课题的主要任务 3
2 图像基础技术与相关概念 4
2.1 数字图像的基本概念 5
2.2 调色板 7
2.3 DDB与 DIB 9
2.4 图像格式的分析、解压与压缩 13
3 踢脚线提取的具体实现 18
3.1 处理类的设计 18
3.2 边缘增强 19
3.3 边缘检测 21
3.3 直线提取 25
4 结论 28
致谢 29
参考文献 30
摘 要
机器人（Robot）是自动执行工作的机器装置。机器人可接受人类指挥，也可以执行预先编排的程序，也可以根据以人工智能技术制定的原则纲领行动。机器人执行的是取代或是协助人类工作的工作，例如制造业、建筑业，或是危险的工作。机器人可以是高级整合控制论、机械电子、计算机、材料和仿生学的产物。目前在工业、医学甚至军事等领域中均有重要用途。
然而机器人导航的任务之一就是避开障碍物。这样，机器人在行走过程中，要充分利用环境中的特征来识别航行区域和障碍物区域。
现在移动机器人的导航方式：惯性导航、视觉导航、基于传感器数据导航、卫星导航等。这些导航方式分别适用于各种不同的环境，包括室内和室外环境，结构化环境与非结构化环境。
本论文阐述了机器人通过视觉传感器在室内特殊的环境中实现自主避障和导航的一种方法。首先讨论了图像及有关图像处理的基本知识,然后通过简单的图像处理方法获取运动环境中的踢脚线。该方法有效地利用了室内环境视觉信息，减少了计算量，提高实时性，通过实践说明了该算法的有效性和可行性。
关键词: 边缘增强、边缘检测、直线提取
Abstract
The Robot is the automatic execution work aggregate. The robot may accept the humanity to direct, may also carry out the procedure which arranges in advance, may also act according to by the artificial intelligence technology formulation principle guiding principle motion. What robot execution is the substitution perhaps assists the work which the humanity works, for example manufacturing industry, architecture industry, perhaps dangerous work. The robot may be the high-level conformity cybernetics, mechanical electronic, the computer, the material and the bionics product. At present in domains and so on industry, medicine even military has the important use.
However one of robot guidance''s duties avoids the obstacle. Thus, the robot in walks in the process, must use in the environment the characteristic to distinguish the navigation area and the obstacle region fully.
Robot''s guidance way now: Inertial navigation, visual guidance, based on sensor data guidance, satellite navigation and so on. These guidance way is suitable separately for each kind of different environment, including indoor and outdoor environment, structurized environment and non-structurized environment.
The present paper elaborated the robot realizes through the visual sensor in the indoor special environment evades independently bonds with the guidance one method. First discussed the image and the related imagery processing elementary knowledge, then kicks the foot line through the simple imagery processing method gain movement environment. This method has used the indoor environment visual information effectively, reduced the computation load, enhances timeliness, explained this algorithm validity and the feasibility through the practice.
Key word: The edge enhancement、The marginal check、The straight line withdraw
1 引言
1.1研究背景
机器人导航的任务之一就是避开障碍物。这样，机器人在行走过程中，要充分利用环境中的特征来识别航行区域和障碍物区域。
现在移动机器人的导航方式：惯性导航、视觉导航、基于传感器数据导航，卫星导航等。这些导航方式分别适用于各种不同的环境，包括室内和室外环境，结构化环境与非结构化环境。
1.2 机器人发展史
1920年 捷克斯洛伐克作家卡雷尔•恰佩克在他的科幻小说《罗萨姆的机器人万能公司》中，根据Robota(捷克文，原意为“劳役、苦工”)和Robotnik(波兰文，原意为“工人”)，创造出“机器人”这个词。
1939年 美国纽约世博会上展出了西屋电气公司制造的家用机器人Elektro。它由电缆控制，可以行走，会说77个字，甚至可以抽烟，不过离真正干家务活还差得远。但它让人们对家用机器人的憧憬变得更加具体。
1942年 美国科幻巨匠阿西莫夫提出“机器人三定律”。虽然这只是科幻小说里的创造，但后来成为学术界默认的研发原则。
1948年 诺伯特•维纳出版《控制论》，阐述了机器中的通信和控制机能与人的神经、感觉机能的共同规律，率先提出以计算机为核心的自动化工厂。
1954年 美国人乔治•德沃尔制造出世界上第一台可编程的机器人，并注册了专利。这种机械手能按照不同的程序从事不同的工作，因此具有通用性和灵活性。
1956年 在达特茅斯会议上，马文•明斯基提出了他对智能机器的看法：智能机器“能够创建周围环境的抽象模型，如果遇到问题，能够从抽象模型中寻找解决方法”。这个定义影响到以后30年智能机器人的研究方向。
1959年 德沃尔与美国发明家约瑟夫•英格伯格联手制造出第一台工业机器人。随后，成立了世界上第一家机器人制造工厂——Unimation公司。由于英格伯格对工业机器人的研发和宣传，他也被称为“工业机器人之父”。
1962年 美国AMF公司生产出“VERSTRAN”(意思是万能搬运),与Unimation公司生产的Unimate一样成为真正商业化的工业机器人，并出口到世界各国，掀起了全世界对机器人和机器人研究的热潮。
1962年-1963年传感器的应用提高了机器人的可操作性。人们试着在机器人上安装各种各样的传感器，包括1961年恩斯特采用的触觉传感器，托莫维奇和博尼1962年在世界上最早的“灵巧手”上用到了压力传感器，而麦卡锡1963年则开始在机器人中加入视觉传感系统，并在1965年，帮助MIT推出了世界上第一个带有视觉传感器，能识别并定位积木的机器人系统。
1968年 美国斯坦福研究所公布他们研发成功的机器人Shakey。它带有视觉传感器，能根据人的指令发现并抓取积木,不过控制它的计算机有一个房间那么大。Shakey可以算是世界第一台智能机器人，拉开了第三代机器人研发的序幕。
1973年 世界上第一次机器人和小型计算机携手合作，就诞生了美国Cincinnati Milacron公司的机器人T3。
1998年 丹麦乐高公司推出机器人(Mind-storms)套件，让机器人制造变得跟搭积木一样，相对简单又能任意拼装，使机器人开始走入个人世界。
1999年 日本索尼公司推出犬型机器人爱宝(AIBO)，当即销售一空，从此娱乐机器人成为目前机器人迈进普通家庭的途径之一。
2002年 丹麦iRobot公司推出了吸尘器机器人Roomba，它能避开障碍，自动设计行进路线，还能在电量不足时，自动驶向充电座。Roomba是目前世界上销量最大、最商业化的家用机器人。
2006年 6月，微软公司推出Microsoft Robotics Studio，机器人模块化、平台统一化的趋势越来越明显，比尔•盖茨预言，家用机器人很快将席卷全球。
1.3机器人视频导航概述
视觉信号具有信号探测范围宽、获取信息丰富等优点。随着近几年图像处理技术以及计算机处理能力、计算机视觉理论及算法的飞速发展，视觉导航成为机器人导航的主要发展方向之一。最近二十年来，基于视觉的室内机器人导航和室外机器人导航均得到了飞速发展。通常，机器人利用装配的摄像机拍摄周围环境的局部图像，然后通过图像处理技术，如特征识别、距离估计等，进行机器人定位及规划下一步的动作。有研究人员利用Fourier变换处理机器人全方位图像，并将关键位置的图像经Fourier变换所得的数据存储起来作为机器人定位的参考点。以后机器人所拍摄的图像经变换后与之相对照，从而得知机器人当前位置。也有研究人员利用视觉技术解决计算机器人运动过程中的避碰点，从而实现机器人局部路径规划。
1.4世界各国机器人研究与应用
美国：美国是机器人的诞生地，早在1962年就研制出世界上第一台工业机器人，比起号称“机器人王国”的日本起步至少要早五六年。经过30多年的发展，美国现已成为世界上的机器人强国之一，基础雄厚，技术先进。
4 结论
记得刚开始选题时，就被机器人视频导航几个字眼深深吸引了，眼球很快了锁定了它，不因为别的，就因为它是人工智能方向的课题。但是随着了解的深入，我后悔当初的选择了，更怀疑自己的眼光。从对C++只是听过到运用写图像处理代码，从图像一无所知到图像文件的解压输入、相关算法的比较，再后来的编程实现，这一步步走来，让我感受很多。以前也写过一些比较大的代码，但都是建立在熟悉编译环境、语言并且有一定理论方面的基础之上。可这次却在一张白纸上进行。开始确实遇到很多困难，有过徘徊与失落。比如说对VC6.0环境的熟悉，刚接触时很想不明白点击菜单中的按纽时到底这个事件是怎样响应并调用相应处理函数的，于是我找了很多相关书籍来阅读，凭借以前的经验我知道最快揭开谜底的办法就是看别人写的代码，看他们在这个地方是怎么实现，在阅读几段代码后，终于搞明白了，其实很简单，系统会直接给你生成！有种众里寻她千百度，那人却在灯火阑珊处的感觉。这与java有很大区别，java需要自己写相关捕获代码。随着深入，又对不同图像文件格式的解压与输入疑惑起来，唯一的办法就是阅读他人的代码，由于不会C++，阅读代码也就更加的困难，时间有限，不可能先学C++再学图像处理，所以我选择了直接学图像处理，每当有看不明白的语法结构时就翻阅C++编程书籍，就这样终于把课题完成。
募然回首，在这次毕业设计中，主要是让我学会了翻阅资料自学，再运用，最后解决问题。其中最重要的是一种方法，一种解决从未涉及领域问题的方法。在以后的学习与工作中遇到问题，不会再因为我没接触过而畏首畏尾，只要为伊消得人憔悴，在回首之时，终有不小的意外。通过做这次毕业设计，我有信心！因为它让我学到了解决此类问题的一套方法！
参考文献
张德慧、周元哲编著;C++ 面向对象程序设计 [M]. 北京:科学出版社,2005.7
崔 屹 编著;数字图像处理技术与应用 [M]. 北京:电子工业出版社,1997.3
周长发编著;精通visual C++ 图像处理编程 [M]. 北京:电子工业出版社,2006.6
董士海等编著;图像格式编程指南 [M]. 北京:清华大学出版社,1994.4
赫荣威编译;计算机图像处理技术 [M]. 北京:北京师范大学出版,1986.1
郑阿奇编著;Visual C++ 教程 [M]. 北京:清华大学出版社,2005.7
Directory
Summary I
ABSTRACT II
1 INTRODUCTION 1
1.1 Background 1
1.2 History of a robot
1.3 Robots Video Navigation Overview 2
1.4 Research and Application of robots around the world 2
1.5 Robot Navigation future direction of three
1.6 The main task of the subject 3
2 image-based technologies and related concepts 4
2.1 The basic concepts of digital image 5
2.2 Palette 7
2.3 DDB and DIB 9
2.4 The analysis of image format decompression and compression 13
3 baseboard extract concrete realization of 18
3.1 dealing with the design of Class 18
3.2 Edge Enhancement 19
3.3 Edge Detection 21
3.3 Line extraction 25
4 Conclusion 28
Thanks 29
References 30
Abstract
Robot (Robot) is automatic implementation of the machine equipment. Robot acceptable human command and can also perform a pre-programmed procedures can also be formulated in accordance with the principles of artificial intelligence program of action. Robots to replace or assist the implementation of the human work, such as manufacturing, construction, or dangerous work. Robot can be a high-level integration of control theory, mechanical electronics, computers, materials and bionic product. Currently in the industrial, medical and even military have important applications in such areas.
Robot navigation, however, one of the tasks is to avoid obstacles. In this way, the process of walking robot, we should take full advantage of the characteristics of the environment to identify areas of navigation and obstacle regions.
Now mobile robot navigation method: inertial navigation, visual navigation, based on sensor data navigation, satellite navigation and so on. These navigation methods were applied to a variety of environments, including indoor and outdoor environment, structured environment and unstructured environment.
This paper described the robot through the vision sensor in the indoor environment to achieve their own unique obstacle avoidance and navigation methods. First discuss the image and the basic knowledge of image processing, and then through a simple image processing method to get the sports environment in the baseboard. This method is effective use of the indoor environment, visual information, reducing the computation and improve real-time, through practice to illustrate the effectiveness of the algorithm and feasibility.
Key words: edge enhancement, edge detection, line extraction
Abstract
The Robot is the automatic execution work aggregate. The robot may accept the humanity to direct, may also carry out the procedure which arranges in advance, may also act according to by the artificial intelligence technology formulation principle guiding principle motion. What robot execution is the substitution perhaps assists the work which the humanity works, for example manufacturing industry, architecture industry, perhaps dangerous work. The robot may be the high-level conformity cybernetics, mechanical electronic, the computer, the material and the bionics product. At present in domains and so on industry, medicine even military has the important use.
However one of robot guidance''s duties avoids the obstacle. Thus, the robot in walks in the process, must use in the environment the characteristic to distinguish the navigation area and the obstacle region fully.
Robot''s guidance way now: Inertial navigation, visual guidance, based on sensor data guidance, satellite navigation and so on. These guidance way is suitable separately for each kind of different environment, including indoor and outdoor environment, structurized environment and non-structurized environment.
The present paper elaborated the robot realizes through the visual sensor in the indoor special environment evades independently bonds with the guidance one method. First discussed the image and the related imagery processing elementary knowledge, then kicks the foot line through the simple imagery processing method gain movement environment. This method has used the indoor environment visual information effectively, reduced the computation load, enhances timeliness, explained this algorithm validity and the feasibility through the practice.
Key word: The edge enhancement, The marginal check, The straight line withdraw
1 Introduction
1.1 Background
Robot navigation is one of the tasks is to avoid obstacles. In this way, the process of walking robot, we should take full advantage of the characteristics of the environment to identify areas of navigation and obstacle regions.
Now mobile robot navigation method: inertial navigation, visual navigation, based on sensor data navigation, satellite navigation and so on. These navigation methods were applied to a variety of environments, including indoor and outdoor environment, structured environment and unstructured environment.
1.2 Robot History
1920 Czechoslovakia writer Karel Capek in his • sci-fi novel "Rossum''s Universal Robots company", according to Robota (Czech, intended to "labor, slave labor") and Robotnik (Polish, the original intent as "workers"), to create a "robot" is the word.
World Expo 1939 in New York on display at Westinghouse Electric Company manufactured home robot Elektro. It is controlled by a cable, you can walk, say 77 words, or even smoke, but still far from the real chores. However, it allows people to household robot''s vision has become more specific.
Asimov sci-fi masters 1942, the United States put forward the "Three Laws of Robotics." Although this is only the creation of science fiction, but later became the principle of academic research and development by default.
• In 1948 Norbert Weiner published in "control theory" to explain the machine in the communication and control function and the nervous, sensory function of the common law, first proposed as the core of computer-automated factory.
1954, American George • Dwyer created the world''s first programmable robot and registered patents. This mechanical hand in accordance with different programs in different jobs, so has the versatility and flexibility.
1956 Dartmouth meeting • Marvin Minsky has made his views on intelligent machines: Smart Machine "to create an abstract model of the surrounding environment, if you encounter problems, from abstract model to find a solution" . This definition affects the subsequent 30 years of intelligent robot research direction.
Dwyer and the United States in 1959, inventor Joseph • Ingeborg joined hands to create the first industrial robot. Subsequently, the establishment of the world''s first a robot manufacturing plant - Unimation company. As Ingeborg R & D for industrial robots and publicity, he was known as the "father of industrial robots."
AMF Inc. in 1962, the United States produced "VERSTRAN" (meaning universal transport), and Unimation produces the same as the Unimate become a true commercial industrial robots, and exported to countries around the world, setting off a worldwide study of robots and robot the globe.
1962 -1,963 years the application of sensors to improve the operability of the robot. People try all kinds of sensors installed on the robot, including the 1961 Ernst used in tactile sensors, Tomovic and Boni 1962, the world''s first "smart hand" on the use of pressure sensors, while the McCarthy in 1963, has begun to add visual sensor in robot system, and in 1965, helped MIT launched the world''s first with a vision sensor that can identify and locate building blocks of the robotic system.
Stanford Research Institute in 1968, the United States announced that they successfully developed a robot Shakey. It is with a vision sensor that can detect and according to the instructions of people crawling the building blocks of a computer to control it, but there is a room so much. Shakey can be regarded as the world''s first intelligent robot, beginning the prelude to the third generation of robot research and development.
In 1973 the world''s first robot and small computers to work together, they gave birth to the U.S. company Cincinnati Milacron robot T3.
In 1998 Denmark introduced Lego Robot (Mind-storms) package, so get with the building-block robot manufacturing the same, relatively simple and can arbitrarily assembled, the robot started to enter a private world.
In 1999 Sony introduced Aibo robot dog (AIBO), immediately sold out, and from entertainment robots become the robot forward one of the ways ordinary family.
In 2002 Denmark introduced the iRobot robotic vacuum cleaner Roomba, it can avoid obstacles, automatic design of the road route, but also in the power is insufficient, automatically towards charging seat. Roomba is the world''s largest-selling and most commercial household robots.
In June 2006, Microsoft launched the Microsoft Robotics Studio, robotics modular, unified platform, it became increasingly evident, Bill • Gates predicted that household robots will soon be sweeping the globe.
1.3 Robots Video Navigation Overview
Visual signals have a wide range of signal detection, access to information-rich and so on. As in recent years, image processing technology and computer processing power, computer vision theory and algorithm of the rapid development of visual navigation as robot navigation, one of the main development direction. The last two decades, based on visual robot navigation in indoor and outdoor robot navigation have been developed rapidly. Typically, the robot assembly of the camera using ambient partial image, and then image processing techniques, such as feature recognition, distance estimation, the robot location and planning the next action. Some studies have used Fourier transform processing robot full image, and the key to the location of the image data obtained by the Fourier transform stored as a reference point for robot localization. The images taken after the robot through the transformation contrast, in order to find robots current position. Some researchers also use visual technology to solve computing robot in the process of collision point, in order to achieve partial robot path planning.
1.4 Research and Application of robots around the world
United States: The United States is the birthplace of the robot, as early as 1962 developed the world''s first industrial robot, compared to known as "robot kingdom" in Japan starting at least as early as 2056. After more than 30 years of development, the United States has become one of the world''s robots power, base, and advanced technology.
4 Conclusion
I remember when the topics of the beginning, he was deeply robot video navigation has attracted a few words, the eye very quickly locked it, not because of anything else, because it is the subject of the direction of artificial intelligence. However, with the in-depth understanding, I have never regretted his choice, and even doubt their own eyes. From the C + + just heard that the use of writing image processing code, no knowledge from image to image file decompression input correlation algorithm comparison, and then the subsequent programming, this step by step approaching, let me feel a lot. Relatively large number of previously written code, but they are compiled based on familiar environment, language, and have some theoretical basis. But this time was carried out in a white paper. The beginning was quite a lot of difficulties, there have been wandering and lost. For example, on the VC6.0 environment, familiar with the new to not understand when you want to click the menu button when in the end is how to respond to this event and call the appropriate handler, and so I find a lot of books to read, by virtue of the previous experience I know that the fastest way to open a mystery is to look at the code written by someone else and see how they are in this place is realized, after reading the paragraphs of the code, and finally engage in to understand, but in fact very simple, the system will be generated directly to you! A kind of public where Sylvia, that people in the happiness is feeling. This java is very different, java need to write your own code related to capture. With the in-depth, but also of different image file formats, extract the input doubts up, the only way is to read other people''s code, because no C + +, read the code will be more difficult, time is limited, can not learn C + + and then the first Science image processing, so I chose the direct learning image processing, whenever there is to see you do not understand the grammatical structure when reading C + + programming books, and thus finally completed the subject.
Had raised Ran Looking back, during the graduation design, mainly to allow the information I have learned to read self-learning, re-use, finally solve the problem. One of the most important is a way, a solution has never been involved in the field of the problem. In the future study and work experience, will not be because I did not come into contact with the over-cautious, as long as consumers get the Iraqi people gaunt, in retrospect, when, and finally there are no small accidents. By doing the graduation project, I am confident! Because it is so I have learned a methodology to solve such problems!
References
Zhang Dehui, Zhou Yuan-zhe ed; C + + object-oriented programming [M]. Beijing: Science Press, 2005.7
CUI Yi ed; digital image processing technology and application [M]. Beijing: Electronic Industry Press, 1997.3
weeks long hair ed; proficient in visual C + + image processing program [M]. Beijing: Electronic Industry Press, 2006.6
Dong Shi-Hai eds; Image Format Programming Guide [M]. Beijing: Tsinghua University Press, 1994.4
He Roewe compile; computer image processing technology [M]. Beijing: Beijing Normal University Publishing, 1986.1
Zheng Achille ed; Visual C + + Tutorial [M]. Beijing: Tsinghua University Press, 2005.7
