Proposed Machine Vision System Research Articles

Vision-Based System for Black Rubber Roller Surface Inspection

This paper proposes a machine vision system for the surface inspection of black rubber rollers in manufacturing processes. The system aims to enhance the surface quality of the rollers by detecting and classifying defects. A lighting system is installed to highlight surface defects. Two algorithms are proposed for defect detection: a traditional-based method and a deep learning-based method. The former is fast but limited to surface defect detection, while the latter is slower but capable of detecting and classifying defects. The accuracy of the algorithms is verified through experiments, with the traditional-based method achieving near-perfect accuracy of approximately 98% for defect detection, and the deep learning-based method achieving an accuracy of approximately 95.2% for defect detection and 96% for defect classification. The proposed machine vision system can significantly improve the surface inspection of black rubber rollers, thereby ensuring high-quality production.

Machine vision system based on a coupled image segmentation algorithm for surface-defect detection of a Si3N4 bearing roller.

Defect detection is a critical way to ensure quality for silicon-nitride-bearing rollers. To improve detection efficiency and precision for silicon-nitride-bearing roller surface defects, in this paper, a novel machine vision system for the detection of its surface defects is designed. This method combines image segmentation and wavelet fusion to extract features from an image. In turn, the features are used in a classifier based on the K-nearest neighbor for defect classification. The optimized image segmentation algorithm that is combined with wavelet fusion is the innovation of the proposed method. It is evaluated using different defect images acquired by the machine vision system. Our experiments show that the proposed machine vision system's precision in anomaly detection of the silicon-nitride-bearing roller surface can achieve 98.5%; further, its classification precision of various defects is greater than 91.5%. It has resulted in a solution for the automatic identification of the silicon-nitride-bearing roller surface defects.

Development of Machine Vision System for Off-Line Inspection of Fine Defects on Glass Screen Surface

The complicated manufacturing process and poor industrial environment inevitably produces unacceptable scratches on glass screen surface. In this study, a machine vision system is developed to automatically detect fine scratch, which has a width smaller than 30 μm, on glass screen surface for the purpose of appearance quality assessment. The imaging hardware of the proposed machine vision system is made up of a high-resolution camera and a wide field lens so that a high-quality image is achieved in a relatively large image area. The captured glass screen images were firstly classified with YOLOv3 to exactly distinguish fine scratches in two other kinds of defects including dusts and stains. Then the U-Net network was applied on the selected scratches to estimate their dimensions approximately. The proposed algorithms combination has a fast image processing speed and it is able to effectively reduce the false segmentation of other defects. The results of the developed machine vision system show a relatively good agreement with the well-established microscopic method. It demonstrates our machine vision method is able to image and quantitative analyze scratches with dimensions between 25 μm to 100 μm.

A Machine Vision System Based on Driving Recorder for Automatic Inspection of Rail Curvature

Because of long distance of railway lines, it is difficult to find an appropriate method to inspect the rail track condition efficiently and accurately. In this article, a machine vision system based on driving recorder and image signal processing is proposed to evaluate the rail curvature automatically. The proposed machine vision system consists of four modules including the video acquisition module, the image extraction module, the image processing module, and the track condition assessment module. Three classic edge detection methods are adopted and compared for rail edge detection. In line with the videos of driving recorder, coordinate systems for train and rail are defined in the Lagrangian space, and the track curvature is estimated using the proposed chord offset method and double measurement method. For evaluating the track condition, an index describing the concordance between the train and track is defined. In the case study, a set of videos from the driving recorders of trains during their in-service operations are analyzed by the proposed technique, and the obtained results are verified by comparison with those obtained by a track geometry inspection vehicle. It is shown that the proposed technique can evaluate the track curvature accurately. Moreover, the influence of the position of deployed driving recorder, the focal length and anti-shake of camera on the accuracy of evaluation results is discussed. It is testified that the proposed technique provides a simple and reliable way to inspect the track curvature.

A Machine Vision Based Automated Quality Control System for Product Dimensional Analysis

Quality control (QC) in manufacturing processes is critical to ensuring consumers receive products with proper functionality and reliability. Faulty products can lead to additional costs for the manufacturer and damage trust in a brand. A growing trend in QC is the use of machine vision (MV) systems because of their noncontact inspection, high repeatability, and relatively low cost. This paper presents a robust MV system developed to perform comparative dimensional inspection on diversely shaped samples, including additive manufacturing products. The algorithm used performs dimensional inspection on a base product considered to have acceptable dimensions. The perimeter, area, rectangularity, and circularity of the base product are determined using blob analysis on a calibrated camera. These parameters are then used as the standard with which to judge additional products. Each product following is similarly inspected and compared to the base product parameters. A likeness score is calculated for each product, which provides a single value tracking all parameter differences. Finally, the likeness score is considered on whether it is within a threshold, and the product is considered to be acceptable or defective. The proposed MV system has achieved satisfactory results, as discussed in the results section, that would allow it to serve as a dependable and accurate QC inspection system in industrial settings.

Infrared imaging based machine vision system to determine transient shape of isotherms in submerged arc welding

Submerged arc welding (SAW) is one of the important industrial welding processes for heavy industries which include ship building and steel industries. The transient temperature distribution during the welding process has a significant influence on the heat affected zone width, microstructure formation, residual stresses, distortion, and hence on the fatigue behaviour of the welded structures. To accurately determine the transient temperature distribution, the exact shape of the heat source distribution needs to be considered. Machine vision can be a unique method for determination of such transient temperature distribution. In this paper, a machine vision system is developed to determine the transient shape of isotherm of submerged arc welded plates. It is based on infrared imaging to capture and extract the temperature history of the welding process. The transient temperature distribution is observed to be egg shape. The shape of isotherm is analyzed and validated with the experimental data acquired for a wide range of process parameters. These comparisons showed close agreement. The extracted data from the proposed machine vision system is used as the input information for various investigations, such as maximum temperature, isotherm generation, cooling rate plots and image segmentation studies. The developed method is a contactless method. It not only eliminates the possible errors encountered in contact measurement methods, but also provides a user friendly interface and accurate results in less time, and hence significantly reduces the experimental time. This method demonstrated its successful application for submerged arc welding, and furthermore it can be applied for the estimation of transient temperature field in other manufacturing processes.

Detection of Foreign Materials in Wheat Kernels using Boundary Descriptors

The present paper reports the development of a machine vision system for quality inspection of wheat using kernel shape attribute. Shape attribute of agricultural products including wheat kernels is extremely difficult to quantify in digital computation. A new method is proposed in the present work to quantify shape attribute of wheat kernels using regional boundary descriptors. Recognition task in the proposed machine vision system is carried out by neural classifier trained with Levenberg-Marquardt (LM) based supervised learning. Proposed neural classifier was executed using feed-forward back-propagation based three layer artificial neural network. Experimental results indicate more than 98.1% overall average classification accuracy for the involved wheat and impurity elements in the present work. The results of present study are quite promising and the proposed machine vision system has potential future for on-line inspection of agriculture produce in real time environment.

Machine vision system for key-cap flatness measurement of computer keyboards

Key-cap flatness detection after assembly is one of the basic quality control (QC) indexes in computer keyboard manufacturing. A modified machine vision system based on linear structured light imaging for measuring the key-cap flatness is proposed for keyboard QC automation. After a brief introduction of the system design and principle, the pipeline of light stripe image processing, especially the removal of printed letter interference, is studied. First, the staggered reprojection of dense multiline fringes is presented using the pattern editability of the digital light processing projector to replace the conventional three-dimensional (3-D) sensor mechanical scanning and avoid the movement and cumulative error. Second, an adaptive direction operator based on a Hough transform voting is proposed. This operator is used for directional morphological filtering to remove letter noise and solve the issue of printed letter interference on the key-cap surface, thus improving the accuracy and stability of stripe centerline extraction. Finally, the nonlinear least square method is used to fit the 3-D surface of the key-cap and evaluate its flatness efficiently based on the discrete globally distributed 3-D point cloud data. The experimental result demonstrates that the proposed machine vision system can quickly detect keyboard key-cap flatness and shows superior performance to that of the previous work.

Contaminated Facade Identification Using Convolutional Neural Network and Image Processing

In recent years, as number of new building getting larger, there has been an increased interest in the cleaning of exterior walls. Accordingly, there is a growing interest in automatic cleaning robots that move around the outer building facade. These robots are also required to apply different cleaning methods to remove various contaminants on the outer wall of the building. However, current surface contaminant detection systems can either detect only a single type of contaminant, or are not compact enough for installation on mobile platforms that move around the outer facade. As cleaning workers are able to distinguish various contaminants with the naked eye, we aim to solve this problem by developing a machine-vision system using convolutional neural networks (CNNs) and image processing methods. As it is a compact system that uses only a camera to take pictures and a processor to process the images, it is suitable for applications involving mobile platforms. Object-type contaminants such as avian feces are handled by the YOLOv3 module using the object-detection algorithm. Area-type contaminants such as rusty stains are processed using the color-detection module using the HSV color space, median filter, and flood fill algorithm. Particle-type contaminants such as dust are handled by the grayscale module, converting images to grayscale images and then comparing the average brightness with a reference that is provided in advance. This proposed machine vision system will detect objects, areas, and particle-type contaminants with a single image and some reference images provided in advance.

Detection of Foreign Materials in Wheat Kernels using Regional Texture Descriptors

The present paper reports the development of an efficient machine vision system for automatic detection of foreign materials in wheat kernels using regional texture descriptors. In this system, the detection task is performed in two phases. These phases include features extraction phase followed by classification phase. New surface texture descriptors of wheat kernels are developed using Non-Shannon entropies in this work. These entropies are defined using intensity histograms of wheat and non-wheat regions in the given image. Such an attempt has not been made earlier. Experimental results on a database of about 2635 wheat and non-wheat components from 63 images confirm the effectiveness of the proposed method. The classification task is performed by the neural classifier in the proposed machine vision system. An accuracy of more than 98.5% is achieved using proposed system. However, the results of present investigations are quite promising.

Robust active vision industrial CAD parts recognition system

In automated assembly systems the machine parts identification is entirely different from simple object recognition; moreover the ability of humans to differentiate between correct and not correct machine parts is better but it is a difficult task for a machine. In general, with fast moving machine parts on the conveyor manual defect detection by human inspectors is impractical. Also it is expensive, inaccurate, subjective, eye straining and causes other health issues to quality control inspectors. A computer vision-based non-contact inspection technique is developed with image processing methods by considering these problems, for defect detection in industrial machine parts. The present work will help the industrial robot used in assembly process and industrial inspection systems. In this paper features-based industrial object detection techniques are implemented in MATLAB to recognise the presence of the industrial CAD parts in the query image. In the end the actual industrial tool images are also used to show the accuracy and robustness of the proposed machine vision system for industrial manufacturing automation.

Vision Based Measurement applied to Industrial Instrumentation

Machine Vision applied in the area of Instrumentation and Measurement is known as Vision Based Measurement. Its application in industry is steadily increasing principally due to the recent developments in computer hardware and computer vision algorithms. In this article we discuss the possibilities of combining industrial local sensors with a camera and a computer, thus obtaining a new type of transmitter based in Machine Vision. The possibility of transforming visual information into digital data, which can be readily available in an industrial network, opens a variety of new applications. As a case study, we propose and implement a Vision Based Measurement system capable of measuring and transmitting the value indicated by a pointer-scale indicator. To verify the applicability of the proposed transmitter the following results are presented: (i) the associated standard uncertainty is evaluated; (ii) the influence of the image size in the transmitter’s uncertainty is investigated; (iii) the proposed transmitter is compared with a regular electronic transmitter and, (iv) its processing time is evaluated. The results obtained in laboratory are promising and indicate that the use of the proposed Machine Vision system in industry is a real possibility.

Machine vision method for non-contact measurement of surface roughness of a rotating workpiece

Purpose – This paper aims to propose a non-contact method using machine vision for measuring the surface roughness of a rotating workpiece at speeds of up to 4,000 rpm. Design/methodology/approach – A commercial digital single-lens-reflex camera with high shutter speed and backlight was used to capture a silhouette of the rotating workpiece profile. The roughness profile was extracted at sub-pixel accuracy from the captured images using the moment invariant method of edge detection. The average (Ra), root-mean square (Rq) and peak-to-valley (Rt) roughness parameters were measured for ten different specimens at spindle speeds of up to 4,000 rpm. The roughness values measured using the proposed machine vision system were verified using the stylus profilometer. Findings – The roughness values measured using the proposed method show high correlation (up to 0.997 for Ra) with those determined using the profilometer. The mean differences in Ra, Rq and Rt between the two methods were only 4.66, 3.29 and 3.70 per cent, respectively. Practical implications – The proposed method has significant potential for application in the in-process roughness measurement and tool condition monitoring from workpiece profile signature during turning, thus, obviating the need to stop the machine. Originality/value – The machine vision method combined with sub-pixel edge detection has not been applied to measure the roughness of a rotating workpiece.

A machine vision system to estimate cotton fiber maturity from longitudinal view using a transfer learning approach

This paper describes a proposed machine vision system developed to acquire longitudinal images of complete cotton fibers and then estimate their average maturity using image and pattern analysis. Maturity is important to the cotton industry because it relates to fiber's propensity to break when submitted to mechanical stress and it influences the quality of the goods produced from it (yarns and fabrics). The proposed system is novel because it estimates maturity indirectly from fibers' longitudinal views using auxiliary training data generated from fibers' cross-sectional views. It uses the transfer learning framework to reconcile the distribution differences between the two views before traditional machine learning algorithms are applied to learn a suitable model for cotton fiber maturity imaged longitudinally. In addition, the proposed system is more descriptive than commercially available systems currently used in the cotton industry because it estimates not only the average maturity of a complete cotton fiber, but also the maturity variations along the fiber from end to end. Validation studies performed show that the transfer learning approach is a practical and promising way to train our system.

Prediction of surface roughness in CNC end milling by machine vision system using artificial neural network based on 2D Fourier transform

This paper presents a system for automated, non-contact, and flexible prediction of surface roughness of end-milled parts through a machine vision system which is integrated with an artificial neural network (ANN). The images of milled surface grabbed by the machine vision system could be extracted using the algorithm developed in this work, in the spatial frequency domain using a two-dimensional Fourier transform to get the features of image texture (major peak frequency F 1, principal component magnitude squared value F 2, and the average gray level G a). Since F1 is the distance between the major peak and the origin, it is a robust measure to overcome the effect of lighting of the environment. The periodically occurring features such as feed marks and tool marks present in the gray-level image can be easily observed from the principal component magnitude squared value F 2. The experimental machining variables speed S, feedrate F, depth of cut D, and the response extracted image variables F 1, F 2, and G a could be used as input data, and the response surface roughness R a measured by Surfcorder SE-1100 (traditional stylus method) could be used as output data of an ANN ability to construct the relationships between input and output variables. The ANN was trained using the back-propagation algorithm developed in this work due to its superior strength in pattern recognition and reasonable speed. Using the trained ANN, the experimental result had shown that the surface roughness of milled parts predicted by machine vision system over a wide range of machining conditions could be got with a reasonable accuracy compared with those measured by traditional stylus method. Compared with the stylus method, the constructed machine vision system is a useful method for prediction of the surface roughness faster, with a lower price, and lower environment noise in manufacturing process. Experimental results have shown that the proposed machine vision system can be implemented for automated prediction of surface roughness with accuracy of 97.53%. The results are encouraging that machine vision system can be extended to many real-time industrial prediction applications.

기계시각장치에 의한 토마토 작물의 병해엽 검출

Machine vision system was used for analyzing leaf color disorders of tomato plants in a greenhouse. From the day when a few leave of tomato plants had started to wither, a series of images were captured by 4 times during 14 days. Among several color image spaces, Saturation frame in HSI color space was adequate to eliminate a background and Hue frame was good to detect infected disease area and tomato fruits. The processed image (<TEX>$G{\sqcup}b^*$</TEX> image) by OR operation between G frame in RGB color space and <TEX>$b^*$</TEX> frame in <TEX>$La^*b^*$</TEX> color space was useful for image segmentation of a plant canopy area. This study calculated a ratio of the infected area to the plant canopy and manually analyzed leaf color disorders through an image segmentation for Hue frame of a tomato plant image. For automatically analyzing plant leave disease, this study selected twenty-seven color patches on the calibration bars as the corresponding to leaf color disorders. These selected color patches could represent 97% of the infected area analyzed by the manual method. Using only ten color patches among twenty-seven ones could represent over 85% of the infected area. This paper showed a proposed machine vision system may be effective for evaluating various leaf color disorders of plants growing in a greenhouse.

머신비전 시스템을 이용한 마이크로드릴 마멸의 상태감시

Micro-drill wear deteriorates accuracy and productivity of the micro components. In order to improve productivity and quality of micro components, it is required to investigate micro-drill wear exactly. In this study, a machine vision system is proposed to measure the wear of micro-drills using a precision servo stage. Calibration experiments are conducted to compensate for the machine vision system. In this paper, worn volume, area and length are defined as wear amounts. Micro-drill wear is reconstructed as the 3D topography and the quantized wear amount by using the shape from focus (SFF) method and wear parameters. Experiments have been conducted with HSS twist micro-drills and SM45C carbon steel workpieces. Validity of the proposed machine vision system is confirmed through experiments.

A machine vision system for measuring the eccentricity of bearings

In this paper, we describe a machine vision system for inspecting bearings, which are an important part of electro-mechanical kWh meters. The system consists of a personal computer with a frame grabber, a black and white progressive scan CCD camera, and a mechanical device with a stepper motor controlled by a special controller. The quality of a bearing depends on the eccentricity of the needle and the length of the needle that extends out of the cylinder. These two features are robustly and accurately determined by the successive application of the Hough transform (HT) and linear regression. The proposed machine vision system allows fast, accurate, reproducible, and robust 100% inspection of bearings of electro-mechanical kWh meters and as such may be a valuable tool for ensuring high-end-product quality.