Yapay Zeka'nın robot görmesi üzerine uygulanması
An Application of robot vision in artificial intelligence
- Tez No: 21863
- Danışmanlar: PROF. DR. TALHA DİNİBÜTÜN
- Tez Türü: Yüksek Lisans
- Konular: Makine Mühendisliği, Mechanical Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 1992
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 90
Özet
ÖZET Bu çalışmada Yapay Zeka yöntemlerinin robot görmesine uygulanması incelenmiş ve ileri aşamadaki çalışmalarda cisim tanıma konusunda çok başarılı sonuçların alınabileceği anlaşılmıştır. Cisim tanıma üzerine yapılan PROLOG programında, kenar çıkarma işlemi sonucunda oluşturulan, zincir kodlan çeşitli işlemlere tabii tutularak yeni bir yöntem geliştirilmiştir. Bu yöntem sayesinde farklı oryantasyonlarda gelen cisimlerin başarıyla tanınması sağlanmıştır. Yazılım ve donanımın geliştirilmesiyle başarının artması mümkündür. iv
Özet (Çeviri)
An Application of Robot Vision in Artificial Intelligence SUMMARY Researchers are interested in developing and designing the next generation of robots than in applying existing robots. That new generation will be characterized by applications of Artificial Intelligence. Present robotics is largely based on control theory cocepts that have come out of engineering. There are various kinds of strong restrictions about classical engineering technology. Because we have been guided by the mathematics and what can be done with the mathematics. The promise of AI is that we will have a number of ways of produsing responses to the environment, responses to stimuli, that are more flexible than the kinds of responses us can get with this classical control technology; whereas with artificial intelligence techniques, we can use natural language input and other kinds of symbolic inputs that are not numbers. An important current bottleneck in how far we can apply robotics is on the side of sensory devices. What kind of information can the robot take in from the environment? So we are interested in releation with robotic vision and Artificial Intelligence. Artificial Intelligence is of growing interdisciplinary interest and practical importance. People with widely varying backgrounds and professions are discovering new ideas and new tools in this young science. Theory-minded psychologists have developed new models of the mind based on the fundamental concepts of AI- symbol systems and information processing. AI is the part of the computer science concerned with designing intelligent computer systems, that is, systems that exhibit the characteristics we associate with intelligence in human behaviour-understanding language, learning, reasoning, solving problems and so on. Many believe that insights into the nature of the mind can be gained by studying that operations of such programs. Since the field first evolved in the mid-1950s. Experimental AI systems have already generated interest and enthusiasm in industry and are being developed commercially. These experimental systems include programs that; 1- solve some hard problems in chemistry, biolgy, geology, engineering, and medicine at human-expert levels of performance, 2- manipulate robotic devices to perform some useful, repetitive, sensory- motor tasks, and 3- answer questions posed in simple dialects of English. There is every indication that useful AI programs will play an importantpart in the evolving role of computers in our lives. Within most scientific disciplines there are several distinct areas of research, each with its own specific interests, research techniques, and terminology. In AI, these specializations include research on language understanding, vision systems, problem solving, AI tools and programming languages, automatic programing, and several others. Most AI research projects are concerned with many, if not all, of these aspects of intelligence. In the first Chapter we introduced aspect of AI. One of them expert systems that are computer programs. Expert systems are constructed to do the kinds of activities that human experts can do, such as design, compose, plan, diagnose, interpret, summarize, audit, give advice. Expert systems occupy a narrow but very important corner of the entire programming establishment. The goal of AI scientists had always been to develop computer programs that could in some sense think, that is, solve problems in a way that would be considered intelligent if done by a human. Expert systems are the fruit of a 20-year quest to define the appropriate nature of such programs. The process of building an expert system is often called knowledge engineering. It typically involves a special form of interaction between the expert-system builder, called the knowledge engineer, and one or more human experts in some problem area. The knowledge engineer“extracts”from the human experts their procedures, strategies, and rules of thumb for problem solving, and builds this knowledge into the expert system. The result is a computer program that solves problems in much the same manner as the human experts. Collection of domain knowledge is called the knowledge base, while the general problem-solving knowledge is called the inference engine. A program with knowledge organized this way is called a knowledge-based system. The knowledge base in expert system contains facts and rules that use those fact as the basis for desicion making. The inference engine contains an interpreter that decides how to apply the rules to infer new knowledge and a scheduler that decides the order in which the rules should be applied. Learning is a procedure in artificial intelligence by which an artificial intelligence program improves its performance by gaining knowledge. Many of the terms used in artificial intelligence to describe learning are borrowed from psychology. There are two basic types of learning-rote learning and cognitive learning. Many of the different types of learning such as learning by discovery, learning by example, learning by analogy, and others can be classed under cognitive learning. In a production system, learning can take place through the automatic acquisition, modification, or deletion of rules. Learning by analogy is ability to recognize the similarity between two problem areas and use rules developed in one problem area to solve a problem in the second problem area. Learning by being told is a computer program that can learn from instructions. Learning by discovery is a type of unsupervised learning in which incoming data are used to form rules so VIthat the system is able to understand the phenomenon under study. Natural language processing is abranch of artificial intelligence programming whose goal is to facilitate communications between humans and computer using written human language. It consists of two areas: natural language understanding and natural language generation. Three important areas of study in natural language processing are lexical analysis, syntactic analysis, and semantic analysis. A computer language that excels in symbol manipulation, pattern matching, and flexibility in constructing knowledge structures. The flexibility in knowledge structures allow AI languages to handle the frequently unpredictable information packet found in Artificial Intelligence applications. Artificial intelligence languages usally follow an applicative or declarative style of programming as opposed to the sequential style of programming found in conventional languages. This style of programming allows the greater flexibility need in dealing with nonalgorithmic problems. The languages that generally fit these criteria include LISP, PROLOG. LISP is LISt Processing language. LISP is based on recursive function theory. It is characterized by symbol manipulations as opposed to numerical manipulation. A declarative language based on logic programming. PROLOG stands for PROgramming in LOGic. PROLOG is based on predicate calculus. The clause is the primary representational scheme in PROLOG and it is used to represent relationships among objects in facts and rules. LISP, PROLOG is faster and does not take up as much memory and has a built-in definite-clause grammar. PROLOG like LISP also uses lists to represent information. Another part of AI research that is receiving increasing attention involves programs that manipulate robot devices. Research in this field has looked at everything from the optimal movement of robot arms to methods of planning a sequence of actions to achieve a robot's goals. Although more complex systems have been built, the thousands of robots that are being used today in industrial applications are simple devices that have been programmed to perform some repetitive task. Most industrial robots are“blind”, but some see through a TV camera that transmits an array of information back to the computer. Processing visual information is another very active, and very difficult, area of AI research. Programs have been devoloped that can recognize objects and shadows in visual scenes, and even identify small changes from one picture to the next. Chapter 3 discusses robot vision. For many tasks that a robot performs, vision is the most important source of information about its environment. The recognition of surronding objects, the perception of certain relations among these objects, that and appropriate responses to a given scene lie at the foundation of much robot activity. Robot vision has been present in varying degrees since the late 1960s. Although there have been remarkable advances since then, the field must still be considered relatively undeveloped. Parts of the problem lies in the fact that the sophistication of the human visual system makes vnrobot vision systems pale by comparision. A rudimentary vision system can augment a robot's capabilities manyfold, and may permit activities that would otherwise be difficult if not impossible to perform. Hardware and software required to gather, process, and act on visual information. Hardware consists of cameras, scanners, preprocessors, computers, and interface equipment. Software includes the algorithms, procedures, and programs required to convert the dijital image into useful information. Input to the hardware is a visual image; the output produced by the software is an action or decision by the robot. The automation of visual processes encompasses several levels of detail and coceptualization, and the sections of this chapter are arranged on this basis. At the lowest level, there must be means of capturing and representing pictures in a robot's memory that is amenable to direct digital computation. Next, certain preprocessing occurs that is designed to enhance useful features of a picture and/or suppress noise and other unwanted aspects. Then, in many applications, edges and lines in the picture are detected. These edges are then synthesized into recognizable objects in the scene. Finally, the objects may relate to each other in a manner that invites a mechanistic“understanding”of a whole scene. The robot can then respond to the interpretation of this scene by carrying out appropriate tasks. The picture is captured, converted into digital form and stored in the robot's memory at the finest level of detail. Two major applications of vision sensing are for the control of manipulators and for automated inspection. To control manipulation, a vision system must be able to identify and locate workpieces and it must be able to determine their orientation. We consider three broad classes of computer vision: image processing; image analysis; and image comprehension. During the past 40 years, there has been a considerable growth of interest in problems of pattern recognition and image processing. Pattern recognition is concerned primarily with the description and classification of measurements taken from physical or mental processes. Although pattern recognition and image processing have developed as two separate disciplines, they are very closely related. The area of image processing consists not only of coding, filtering, enhancement, and restoration, but also analysis and recognition of images. On the other hand, the area of pattern recognition includes not only feature extraction and classification, but also preprocessing and description of patterns. A common task in computer vision is to recognize the objects in an image. Most computer vision systems do this by matching models for each vmpossible object type in turn, recognizing objects by the best matches. This is not ideal, as it does not take advantages of the similarities and differences between the possible object types. The computation time also increases linearly with the number of possible objects, which can become a problem if the number is large. A new recognition method is described: feature indexed hypotheses, which takes advantage of the similarities and differences between object types and is able to handle cases, where there are a large number of possible object types, in sublinear computation time. A two-dimensional occluded parts recognition system using feature indexed hypotheses is described. IX
Benzer Tezler
- COLREGs-compliant and non-prioritized motion planning for autonomous unmanned surface vehicles
Otonom insansız deniz araçları için COLREG-uyumlu ve önceliksiz hareket planlaması
MUSTAFA BAYRAK
Yüksek Lisans
İngilizce
2023
Elektrik ve Elektronik Mühendisliğiİstanbul Medeniyet ÜniversitesiElektrik-Elektronik Mühendisliği Ana Bilim Dalı
DR. ÖĞR. ÜYESİ HALUK BAYRAM
- Zeki etmenlerde öğrenme kabiliyetinin geliştirilmesi ve dinamik bir atölye tipi çizelgeleme uygulaması
Başlık çevirisi yok
MEHMET EMİN AYDIN
Doktora
Türkçe
1997
Endüstri ve Endüstri MühendisliğiSakarya ÜniversitesiEndüstri Mühendisliği Ana Bilim Dalı
DOÇ.DR. ERCAN ÖZTEMEL
- Soğuk Savaştan akıllı savaşa geçiş sürecinde yapay zekâ teknolojilerinin rolü: Çin ve ABD örnekleri
Role of artifical intelligence techologies in transition from Cold War to intelligent war: Examples of China and the United States
ORKUN ÖZ
Yüksek Lisans
Türkçe
2023
Kamu YönetimiKapadokya ÜniversitesiSiyaset Bilimi ve Kamu Yönetimi Ana Bilim Dalı
DR. ÖĞR. ÜYESİ HANDE EREN
- Bağışıklık sistemleri kullanılarak çoklu etmen sistemlerinde öğrenme
Learning in multi-agent system with artificial immune system
SEMİYE YILMAZ
Yüksek Lisans
Türkçe
2006
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve KontrolSelçuk ÜniversitesiPROF. DR. AHMET ARSLAN
- Generative model-based control strategies for soft robotics arms
Yumuşak robotik kollar için üretiken model tabanlı kontrol stratejileri
ABDELRAHMAN ALKHODARY
Doktora
İngilizce
2023
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve KontrolBahçeşehir ÜniversitesiBilgisayar Mühendisliği Ana Bilim Dalı
Assoc. Prof. Dr. MEHMET BERKE GÜR