There is a trend in manufacturing towards fully automated production facilities in which all operations are integrated by computer based information systems. The current generation of industrial inspection systems lack the necessary flexibility to operate in these environments. AI based Image Understanding systems have the necessary level of generality, achieved through the use of domain specific object models. These models are used to guide early visual processing, and must be supplied to the system. Current theories in cognitive psychology call for a reevaluation of the role of such 'auxiliary' knowledge in early visual processing. Recent work suggests that very general cognitive processes may build up a hierarchical representation of the world. The emphasis is currently on such generic cognitive processes rather than on the use of world knowledge. A novel approach to image processing, in which emphasis is placed on generic low and intermediate level techniques, is proposed in this thesis. This approach, termed the descriptor approach, delays the use of domain specific models until a full description of the image has been produced. A prototype industrial inspection system has been implemented, based on the descriptor approach: the Hierarchical Scene Description (HISD) system. General image features are extracted from images of populated PCBs, and subsequently transformed into a database of prolog facts by an interface subsystem. Finally the intermediate level vision subsystem uses rules to reason about these features, building up a semantic net based description of the scene. HISD successfully builds up hierarchical descriptions of real industrial PCB images in terms of geometric shapes, their coordinates, and spatial relationships between shapes. The results are displayed graphically and are achieved without the use of any object models, thus avoiding the problems of inflexibility and lack of generality associated with more complex model based systems.

Neste trabalho é proposto a técnica da Transformada de Hough para a Inspeção Visual Automática de placas retangulares, visando a determinação de características tais como perímetro, ortogonalidade, centro de massa, área, independente de rotação e translação do objeto dentro da imagem. Os resultados obtidos por essa técnica são avaliados pela comparação com o método dos Momentos, o recurso mais utilizado para a determinação de área e centro de massa para imagens bi-dimensionais. A análise estatística dos dados mostra a robustez da técnica da Transformada Hough para imagens ruidosas. / In this work, is proposed the Hough transform method for the Automated Visual lnspection of rectangular boards. The aim is to extract features such as position, perimeter and area, despite of object translation and rotation. The results obtained by this method are evaluated by comparing them with the method of Moments, the most used method to determine the area and center of mass for bi-dimensional images. Statistical analysis show the robustness of Hough transform for noisy images visual inspection.

Neste trabalho é proposto a técnica da Transformada de Hough para a Inspeção Visual Automática de placas retangulares, visando a determinação de características tais como perímetro, ortogonalidade, centro de massa, área, independente de rotação e translação do objeto dentro da imagem. Os resultados obtidos por essa técnica são avaliados pela comparação com o método dos Momentos, o recurso mais utilizado para a determinação de área e centro de massa para imagens bi-dimensionais. A análise estatística dos dados mostra a robustez da técnica da Transformada Hough para imagens ruidosas. / In this work, is proposed the Hough transform method for the Automated Visual lnspection of rectangular boards. The aim is to extract features such as position, perimeter and area, despite of object translation and rotation. The results obtained by this method are evaluated by comparing them with the method of Moments, the most used method to determine the area and center of mass for bi-dimensional images. Statistical analysis show the robustness of Hough transform for noisy images visual inspection.

Automated visual inspection is an emerging multi-disciplinary field with many challenges; it combines different aspects of computer vision, pattern recognition, automation, and control systems. There does not exist a large body of work dedicated to the design of generalized visual inspection systems; that is, those that might easily be made applicable to different product types. This is an important oversight, in that many improvements in design and implementation times, as well as costs, might be realized with a system that could easily be made to function in different production environments. <br /><br /> This thesis proposes a framework for generalizing and automating the design of the defect classification stage of an automated visual inspection system. It involves using an expandable set of features which are optimized along with the classifier operating on them in order to adapt to the application at hand. The particular implementation explored involves optimizing the feature set in disjoint sets logically grouped by feature type to keep search spaces reasonable. Operator input is kept at a minimum throughout this customization process, since it is limited only to those cases in which the existing feature library cannot adequately delineate the classes at hand, at which time new features (or pools) may have to be introduced by an engineer with experience in the domain. <br /><br /> Two novel methods are put forward which fit well within this framework: cluster-space and hybrid-space classifiers. They are compared in a series of tests against both standard benchmark classifiers, as well as mean and majority vote multi-classifiers, on feature sets comprised of just the logical feature subsets, as well as the entire feature sets formed by their union. The proposed classifiers as well as the benchmarks are optimized with both a progressive combinatorial approach and with an genetic algorithm. Experimentation was performed on true colour industrial lumber defect images, as well as binary hand-written digits. <br /><br /> Based on the experiments conducted in this work, it was found that the sequentially optimized multi hybrid-space methods are capable of matching the performances of the benchmark classifiers on the lumber data, with the exception of the mean-rule multi-classifiers, which dominated most experiments by approximately 3% in classification accuracy. The genetic algorithm optimized hybrid-space multi-classifier achieved best performance however; an accuracy of 79. 2%. <br /><br /> The numeral dataset results were less promising; the proposed methods could not equal benchmark performance. This is probably because the numeral feature-sets were much more conducive to good class separation, with standard benchmark accuracies approaching 95% not uncommon. This indicates that the cluster-space transform inherent to the proposed methods appear to be most useful in highly dependant or confusing feature-spaces, a hypothesis supported by the outstanding performance of the single hybrid-space classifier in the difficult texture feature subspace: 42. 6% accuracy, a 6% increase over the best benchmark performance. <br /><br /> The generalized framework proposed appears promising, because classifier performance over feature sets formed by the union of independently optimized feature subsets regularly met and exceeded those classifiers operating on feature sets formed by the optimization of the feature set in its entirety. This finding corroborates earlier work with similar results [3, 9], and is an aspect of pattern recognition that should be examined further.

Automated visual inspection is an emerging multi-disciplinary field with many challenges; it combines different aspects of computer vision, pattern recognition, automation, and control systems. There does not exist a large body of work dedicated to the design of generalized visual inspection systems; that is, those that might easily be made applicable to different product types. This is an important oversight, in that many improvements in design and implementation times, as well as costs, might be realized with a system that could easily be made to function in different production environments. <br /><br /> This thesis proposes a framework for generalizing and automating the design of the defect classification stage of an automated visual inspection system. It involves using an expandable set of features which are optimized along with the classifier operating on them in order to adapt to the application at hand. The particular implementation explored involves optimizing the feature set in disjoint sets logically grouped by feature type to keep search spaces reasonable. Operator input is kept at a minimum throughout this customization process, since it is limited only to those cases in which the existing feature library cannot adequately delineate the classes at hand, at which time new features (or pools) may have to be introduced by an engineer with experience in the domain. <br /><br /> Two novel methods are put forward which fit well within this framework: cluster-space and hybrid-space classifiers. They are compared in a series of tests against both standard benchmark classifiers, as well as mean and majority vote multi-classifiers, on feature sets comprised of just the logical feature subsets, as well as the entire feature sets formed by their union. The proposed classifiers as well as the benchmarks are optimized with both a progressive combinatorial approach and with an genetic algorithm. Experimentation was performed on true colour industrial lumber defect images, as well as binary hand-written digits. <br /><br /> Based on the experiments conducted in this work, it was found that the sequentially optimized multi hybrid-space methods are capable of matching the performances of the benchmark classifiers on the lumber data, with the exception of the mean-rule multi-classifiers, which dominated most experiments by approximately 3% in classification accuracy. The genetic algorithm optimized hybrid-space multi-classifier achieved best performance however; an accuracy of 79. 2%. <br /><br /> The numeral dataset results were less promising; the proposed methods could not equal benchmark performance. This is probably because the numeral feature-sets were much more conducive to good class separation, with standard benchmark accuracies approaching 95% not uncommon. This indicates that the cluster-space transform inherent to the proposed methods appear to be most useful in highly dependant or confusing feature-spaces, a hypothesis supported by the outstanding performance of the single hybrid-space classifier in the difficult texture feature subspace: 42. 6% accuracy, a 6% increase over the best benchmark performance. <br /><br /> The generalized framework proposed appears promising, because classifier performance over feature sets formed by the union of independently optimized feature subsets regularly met and exceeded those classifiers operating on feature sets formed by the optimization of the feature set in its entirety. This finding corroborates earlier work with similar results [3, 9], and is an aspect of pattern recognition that should be examined further.