應用視覺於TFT-LCD陣列電路工程中瑕疵辨識之研究-畫素電極光罩瑕疵自動分類

PurposeHeritage is the latent part of a sustainable built environment. Conservation and preservation of heritage is one of the United Nations' (UN) sustainable development goals. Many social and natural factors seriously threaten heritage structures by deteriorating and damaging the original. Therefore, regular visual inspection of heritage structures is necessary for their conservation and preservation. Conventional inspection practice relies on manual inspection, which takes more time and human resources. The inspection system seeks an innovative approach that should be cheaper, faster, safer and less prone to human error than manual inspection. Therefore, this study aims to develop an automatic system of visual inspection for the built heritage.Design/methodology/approachThe artificial intelligence-based automatic defect detection system is developed using the faster R-CNN (faster region-based convolutional neural network) model of object detection to build an automatic visual inspection system. From the English and Dutch cemeteries of Surat (India), images of heritage structures were captured by digital camera to prepare the image data set. This image data set was used for training, validation and testing to develop the automatic defect detection model. While validating this model, its optimum detection accuracy is recorded as 91.58% to detect three types of defects: “spalling,” “exposed bricks” and “cracks.”FindingsThis study develops the model of automatic web-based visual inspection systems for the heritage structures using the faster R-CNN. Then it demonstrates detection of defects of spalling, exposed bricks and cracks existing in the heritage structures. Comparison of conventional (manual) and developed automatic inspection systems reveals that the developed automatic system requires less time and staff. Therefore, the routine inspection can be faster, cheaper, safer and more accurate than the conventional inspection method.Practical implicationsThe study presented here can improve inspecting the built heritages by reducing inspection time and cost, eliminating chances of human errors and accidents and having accurate and consistent information. This study attempts to ensure the sustainability of the built heritage.Originality/valueFor ensuring the sustainability of built heritage, this study presents the artificial intelligence-based methodology for the development of an automatic visual inspection system. The automatic web-based visual inspection system for the built heritage has not been reported in previous studies so far.

The Sixth International Symposium on Neural Networks (ISNN 2009)

Performance evaluation of an automatic inspection system of weld defects in radiographic images based on neuro-classifiers

Design of a glue dispenser route inspection system based upon the track of adhesive glue is the focus of this article. The defects of the glue track such as deformation, offset, scrape, and broken glue may affect the quality of production and efficiency. An automatic dispenser route inspection system in combination with the techniques of back-propagation neural (BPN) network with computer vision is developed. Before dispensing, the positioning process of the dispenser system is significant. A simple positioning method is developed to ensure the glued object mounted on a platform is in an acceptable position for glue shooting so that any likely failure due to inaccurate positioning is avoided. Thus, the positioning problem will therefore not influence the cause-and-effect failure investigation of other factors. The images of the track are acquired and then preprocessed to extract the features (coordinates of edge) for inspections. By checking the number of the searched pixels of the boundary of the glue track compared to the edge number of a uniform one, serious failure can be identified. For further diagnosis, six parameters including the average width and its standard deviation (SD) of the track, average offset and its SD, and the average deviation between the neighboring points on the left and right sides are designed as the input units in the input layer of a three-layer neural network and trained with experimental patterns. Using this BPN network system, the recognition rate is able to achieve 96.45% for additional arbitrarily chosen samples.

Relevance. The study examines the possibility of using a mathematic model of an artificial neural network (ANN) for comparative quantitative morphological assessment of revascularized autograft remodeling after reconstructive surgery in patients with jaw defects.Materials and methods. Jawbone biopsies of 30 patients were histologically examined during the implant placement stage, 6, 9, 12 months after the reconstructive surgery with a revascularized fibular graft. Histology images were analyzed using a mathematic model created on the basis of GoogLeNet and trained on morphological images of bone maturation. Histology slides were digitized by a scanning microscope Leica SC2 and streamed through the neural network model. Results. During the follow-up period, jaw defect areas demonstrated relatively mature bone tissue formation with various intensity rates of remodeling and maturation of the newly-formed bone. At that, the results of the descriptive histology were consistent with the quantitative results of the ANN mathematic model, created on the basis of the softwarehardware system developed by the authors. The confirmed significance level is 95% or higher.Conclusion. Pathomorphological data of biopsies were studied using an ANN mathematical model, built on a softwarehardware system, which allowed analyzing all microscope fields of view of a histological slide, bypassing random samples, as well as the evaluation by a pathologist of individual microscope fields of view, to exclude the possibility of unrepresentative sampling and the influence of human factor, which significantly increases the significance of the received results.

Pipelines are exposed to the severe threat of natural disasters, where the damage caused by landslides are particularly bad. Hence, in the route arrangement and maintenance management of pipeline projects, it is particularly important to evaluate the regional landslide hazards in advance. However, most models are based on the subjective determination of evaluation factors and index weights; this study establishes a quantitative hazard assessment model based on the location of historical landslides and the Levenberg–Marquardt Back Propagation (LM-BP) Neural Network model was applied to the pipeline area. We established an evaluation index system by analyzing the spatial patterns of single assessment factors and the mechanism of landslides. Then, different from previous studies, we built the standard sample matrix of the LM-BP neural network by using interpolation theory to avoid the serious influence of human factors on the hazard assessment. Finally, we used the standard sample matrix and the historical data to learn, train, test, and simulate future results. Our results showed 33 slopes with low hazard (accounting for 10.48% of the total number of slopes and corresponding to approximately 32.63 km2), 62 slopes with moderate hazard (accounting for 19.68% of the total number of slopes and corresponding to approximately 65.53 km2), 112 slopes with high hazard (accounting for 35.56% of the total number of slopes and corresponding to approximately 123.55 km2), and 108 slopes with extremely high hazard (accounting for 34.29% of the total number of slopes and corresponding to approximately 150.65 km2). Local spatial autocorrelation analysis indicated that there are significant “high–high” and “low–low” aggregation of landslide hazards in the pipeline area. By comparing the model results with the past landslides, new landslides and landslide potential points, its prediction capability and accuracy were confirmed. On the basis of the results, our study has developed effective risk prevention and mitigation strategies in mountain areas to promote pipeline safety.

컴퓨터용 Lead Screw는 검사방식이 수동검사로 이루어져 작업자의 주관적 판단으로 인한 미검출과 과검출이 발생하고, 생산성이 저하되는 단점이 있다. 따라서 본 연구에서는 이러한 단점을 보완하고자 Lead Screw 자동검사 시스템을 개발했다. 검사시스템은 Lead Screw 자동공급 및 이송을 위한 공급 이송 장치와, 표면검사를 위한 <TEX>$360^{\circ}$</TEX> 회전장치, 치수검사를 위한 머신비젼 부분으로 구성된다. Lead Screw 자동검사 시스템으로 표면과 치수 측정결과, 오검률 및 미검률이 수동검사 결과보다 우수함을 보였다. In manual inspection of Lead Screw of computers many workers are needed to inspect samples, and its main disadvantage is that such types of inspection system not only gives low production, but also gives low perfection. Besides, in manual inspection system, the inspection cost of samples is higher than that of the automatic inspection system. Therefore, in this study to compensate these shortcomings, an automatic inspection system is developed. For the inspection of the surfaces and different dimensional parameters of computer Lead screw, a <TEX>$360^{\circ}$</TEX> rotating machine vision system is developed. From the detailed analysis of the inspection results using the present developed inspection system, it is observed that the developed Lead Screw automatic inspection system is superior to those of manually inspection system.

As automatic surface inspection systems evolve for metal surfaces, factors affecting defect detection need to be identified and addressed. At Rautaruukki New Technology a system based on a special illumination unit and a CCD-line scan camera has been developed. The Smartvis system uses a novel image compression technique and an adaptive defect detection method in digital image processing. It also features a high performance defect classification method including knowledge extraction techniques and fuzzy set theory based algorithms. The improvement of man-machine interactivity has been one of the major goals in the development. In this paper the factors affecting the decisions made in system design are discussed. Experiences from system applications for steel strip process lines are presented.

For defense applications, rapid X-ray inspection of propellant samples is essential for the identification and assessment of defects. Automation of this process using artificial intelligence is possible by properly training a neural network model. Convolution Neural Networks (CNNs) have recently demonstrated excellent success in both the tasks of image recognition and localisation using an adequate amount of data. In real-world, it’s not an easy task to produce the correct amount of experimental data required for the deep neural network to operate. In this work, we propose a method for producing synthetic radiographic data that is supported by ray tracing based radiographic simulations for the deep learning algorithms to automatically detect anomaly in X-ray images. The simulation results, which are then supplemented by noise extracted from the experimental data, show a good comparison with the measurements. This Simulation assisted Automatic Defect Recognition (Sim-ADR) system simultaneously perform defect detection and defect instance segmentation. The accuracy of the defect detection system is more than 87% on a testing set included 416 images.

Electronic Toll Collection (ETC) system is an important part of modern intelligent transport system, which plays an important role in intelligent car parks and congestion charging. Electronic Toll Collection (ETC) system includes automatic vehicle identification system and electronic billing system. The automatic vehicle identification system includes three parts: automatic vehicle identification (AVI), automatic vehicle model identification (AVC) and inspection system (ES).The application of ETC in intelligent car parks and congestion charging also includes automatic vehicle identification system (automatic vehicle identification, automatic vehicle model identification and inspection system) and electronic billing system. In intelligent car parks, ETC technology achieves fast and accurate automatic toll payment by automatically identifying the ETC tags carried by vehicles, which greatly improves the efficiency of car parks and reduces the queuing and waiting time due to manual toll collection. In terms of congestion charging, ETC technology provides an effective means to implement congestion charge collection. By setting up ETC toll lanes in core urban areas or busy road sections, the system can automatically record the information of the vehicles passing through and carry out toll settlement, thus realising dynamic toll collection according to the degree of congestion and the frequency of vehicle use. This congestion charging strategy aims to regulate urban traffic flow through economic means, alleviate traffic congestion and promote the sustainable development of urban transport. Although China's theoretical research on congestion charging, has been relatively mature, but affected by a variety of factors, no city has yet to start the implementation of congestion charging, the lack of experience in practice. Singapore will not stop the car charging system (ETC) technology successfully applied to urban congestion charging and intelligent car parks, ETC system in intelligent car parks and congestion charging application, help to promote the development of urban transport to a more intelligent, green direction. China can learn from the successful experience of foreign countries and apply it to its own intelligent car parks and congestion charging. This paper introduces the basic principle of ETC, discusses the research progress of ETC in terms of technical characteristics, and analyses the application scenarios of ETC in intelligent car parks and congestion charging. By examining the opportunities offered by Electronic Toll Collection (ETC) systems in smart car parks and congestion pricing, this study aims to delve into the transformative impacts of Electronic Toll Collection (ETC) systems on intelligent mobility and contribute to the sustainable development of intelligent mobility.

Digital image processing techniques and advancements help in performing the process of defect detection in hard reflective metal surfaces. Automatic inspection systems have the potential significance to improve quality and increase the production rate. The longitudinal cracks are common defects of continuous casting and may lead to serious quality issues. Image capturing and recognition of hot casting is an effective way detect cracks, and recognition of cracks is essential because the surface of the hot casting is very complicated. To detect the surface cracks of the casting, the feature extraction method is used. The current research work uses multiclass support vector machine (multiclass SVM) technique one of the popular classifiers in image classification. The acquired image is preprocessed by removing the noise present in the image. K-means clustering is applied in segmenting the images based on the specified features, such as size, texture, contrast, etc. The proposed method shows enhanced results in classifying the defective region in the metal surface.

At present, the method of identifying the fault symptoms of various machines by combining the neural network and the D-S evidence theory is attracting attention from researchers because the identification time is fast and the diagnosis is accurate. In this paper, it was mentioned a method for identifying the fault symptoms of automatic transmission by combining these two theories. First, it was mentioned a method for identifying fault symptoms of the automatic transmission by combining a fuzzy neural network and an RBF neural network. Next, it was newly described a method to improve the accuracy of fault symptom identification by the D-S evidence theory. In addition, the accuracy of this method was verified by an experimental method. In the experiment Firstly, two sub neural networks are established to recognize the initial symptoms. That is, the first sub-neural network E1 be used as the fuzzy neural network, the second sub-neural network E2 be used as RBF neural network, respectively, for preliminary symptom recognition. And then, these outputs of the two sub neural networks are used as the evidence space of D-S evidence theory, so the global diagnosis is carried out. The results show that the test results are consistent with the actual fault symptoms. The success rate of fault diagnosis up to 96.3%, therefore, on the identification of the automatic transmission fault symptom, effectiveness, and feasibility of the D-S evidence theory based on information fusion is verified.

The past decade has witnessed an impetus in the rise of artificial intelligence techniques in a diverse range of computer vision-relevant applications. Specifically, automated optical inspection technology has been garnered substantial research attention due to its advantages of enabling amelioration in the manufacturing process and improving the efficiency of industrial production. To date, the studies and evaluations into the defects on the raw leather pieces involving observational analysis tend to rely heavily on domain experts who have undergone intensive training. Herein, an automatic inspection system is introduced to distinguish whether a leather patch contains any defective areas. In particular, the dataset involved in this experiment comprises a total of 375 images that are made up of three classes, viz, black line defect, wrinkle defect and non-defective. Notably, owing to the data scarcity issue, the Generative Adversarial Network (GAN) is adopted to discover the feature regularities to produce plausible additional training samples. The proposed algorithm is systematically analysed, and compelling performance is yielded when compared to previous works. Succinctly, with a relatively small amount of readily captured training data, the classification performance is capable of achieving 100% accuracy. As such, the proposed system may be served an appealing strategy, especially in dealing with leather defect recognition tasks.

Battelle scientists are involved in the development of a real-time multispectral imaging and classification system which can be taken into the field to support automated target/background discrimination. The system also has applications in environmental remote sensing, industrial inspection and medical imaging. The Battelle-developed system consists of a passive, multispectral imaging Electro-Optical (E-O) sensor suite and a real-time digital data collection and data fusion image processor. The E-O sensor suite, able to collect imagery in 12 distinct wavebands from the ultraviolet (UV) through the long wave infrared (LWIR), consists of five charge-coupled device (CCD) cameras and two thermal IR imagers integrated on a common platform. The data collection and processing system consists of video switchers, recorders and a real-time sensor fusion/classification hardware system which takes any three wavebands as input and fuses the bands together by applying look-up tables, derived from tailored neural network algorithms, to classify the scene. The result is then visualized in a video format on a full color, 9-inch, active matrix Liquid Crystal Display (LCD). A variety of classification algorithms including artificial neural networks and data clustering techniques were successfully optimized to perform pixel-level classification of imagery in complex scenes. The optimized classification algorithm is used to populate the look-up tables in the real-time sensor fusion board for use in the field. >

This work proposes a portable, handheld electronic device, which measures the cleanliness in fiber optic connectors via digital image processing and artificial neural networks. Its purpose is to reduce the evaluation subjectivity in visual inspection done by human experts. Although devices with this purpose already exist, they tend to be cost-prohibitive and do not take advantage of neither image processing nor artificial intelligence to improve their results. The device consists of an optical microscope for fiber optic connector analysis, a digital camera adapter, a reduced-board computer, an image processing algorithm, a neural network algorithm and an LCD screen for equipment operation and results visualization. The image processing algorithm applies grayscale histogram equalization, Gaussian filtering, Canny filtering, Hough transform, region of interest segmentation and obtaining radiometric descriptors as inputs to the neural network. Validation consisted of comparing the results by the proposed device with those obtained by agreeing human experts via visual inspection. Results yield an average Cohen's Kappa of 0.926, which implies a very satisfactory performance by the proposed device.