Inspection Of Metallic Surfaces Research Articles

Advancing Crack Detection Using Deep Learning Solutions for Automated Inspection of Metallic Surfaces

The deep Convolutional Neural Network (CNN) architecture used in this research study provides a proof of concept for crack detection on the metallic surface of a hex nut. The goal is to create an automated receiving inspection process to supplement human inspections conducted on-site. Conventional image processing techniques (IPTs) have been extensively used for mechanical infrastructure fault detection. These techniques focus on image modification to extract typical features, such as surface fractures in materials like steel and concrete. However, obstacles presented by a variety of real-world variables, such as changes in lighting and shadows, make it difficult to use IPTs. Our suggested vision-based method employs a deep learning CNN to overcome these difficulties, eliminating the need to explicitly compare fault features. CNNs are more resilient to shifting real-world situations than IPTs since they are naturally trained to identify characteristics in images. Following training on a dataset of 1081 images with dimensions of 256 x 256 pixels, the VGG16 CNN architecture achieved an impressive accuracy of around 94.17%. Additional CNN architectures, including ResNet, MobileNet, AlexNet, and LeNet-5, are employed to assess and compare fault detection accuracies in order to select the most appropriate architecture for the model. To evaluate the robustness and flexibility of the suggested method in various situations, we conducted tests with 206 images from an alternative structure that was not part of the training dataset. These images depicted a range of circumstances, such as intense light patches and tiny fissures. The outcomes showed that our proposed method outperforms current approaches, highlighting its usefulness in practical situations involving the identification of metallic defects.

An Autonomous Robotic Platform for Manipulation and Inspection of Metallic Surfaces in Industry 4.0

Quality control in industry involves trained operators to manipulate and inspect metallic surfaces in order to identify, and eventually correct, manufacturing defects. These tasks are manually performed, and a poor performance ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g.</i> , missing defects) leads to an increase of the costs and prolongation of the manufacturing time cycle. In this work, we propose a multi-agent robotic platform to autonomously perform Industry 4.0 quality control processes of metallic surfaces. The platform consists of three anthropomorphic robots with custom-made end-effectors designed to manipulate, inspect, and eventually correct a metallic frame of a motorcycle. The description of a novel multi-agent platform is followed by the presentation of the developed inspection procedure, in which a linear laser scanner is used to reconstruct the three-dimensional metallic surface of a motorcycle with a resolution of ~0.1 mm. In order to validate the platform, we perform a set of experiments to assess the performance of the robotic platform in a real Industry 4.0 scenario. Results confirmed that such a system guarantees a sub-millimetric precision to identify defects on complex-shaped metallic surfaces and effectively correct them. The proposed robotic platform can be adopted for overcoming the drawbacks of a traditional procedure that relies on visual-tactile manual defects correction ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g.</i> , low-repeatability, high-subjectivity) and is scalable to different industrial applications. The proposed approach aims to elevate the role of operators to expert supervisors of the process, limiting the interactions with potentially-dangerous tools/procedures and thus improving the working conditions in an industrial 4.0 scenario. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —This work was motivated by a crucial need in industry, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e.</i> to automatize the manufacturing quality control, translating the commonly-used visual-based manual approach performed by operators to an objective robotic one that relies on defect detection by using a linear laser scanner. A novel multi-agent robotic platform, developed by the authors, showed its effectiveness in automatizing complex tasks, in which huge workspace and different tools are required. The aim of the paper was to develop a fully automatized application that covers the entire quality control process, while focusing on one specific phase, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e.</i> automatic inspection. The integrated software and the infrastructural communication protocols of the entire robotic platform were designed to be flexible in order to realize a new reference for industrial applications, where a multi-agent approach is demanded. The experimental validation focused on a specific use case (a motorcycle frame selected due to its complex structure, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e.</i> multiple curvatures with variable radii, and large volumes), but the proposed platform and the implemented methodology have to be intended as a general purpose approach, adaptable to any industrial process and mechanical component. The authors, starting by a laboratory development with extensive tests, applied and demonstrated the feasibility of the proposed approach in a real Industry 4.0 scenario.

Using artificial intelligence strategies for process-related automated inspection in the production environment

In this paper a new method for the automatic visual inspection of metallic surfaces is proposed by using Convolutional Neural Networks (CNN). The different combinations of network parameters were developed and tested. The obtained results of CNN were analysed and compared with the results of our previous investigations with color and texture features as input parameters for a Support Vector Machine. Advantages and disadvantages of the different classifying methods are explained.

Theoretical and Experimental Analysis of an Induction Planar Actuator with Different Secondaries--A Planar Driver Application for Metallic Surface Inspection.

This paper presents a study on an induction planar actuator concept. The device uses the same principles as a linear induction motor in which the interaction between a travelling magnetic field and a conducting surface produces eddy currents that leads to the generation of a thrust force and can result in movement over a metallic surface. This can benefit the inspection of metallic surfaces based on the driving platform provided by the induction planar actuator. Equations of the magnetic and electric fields are presented and, by means of these equations, the forces involved were calculated. The behaviour of thrust and normal forces was analysed through the equations and by numerical models, and compared with the results obtained by measurements on a device prototype built in the laboratory as part of the study. With relation to the surface under inspection that forms the secondary, three cases were analysed: (1) a double-layered secondary formed by aluminium and ferromagnetic slabs; (2) a single aluminium layer and (3) a single ferromagnetic layer. Theoretical and measured values of thrust and normal forces showed good correlation.

이동평균 영상처리기법을 이용한 금속판재 표면품질 정밀 측정시스템 연구

산업자동화를 위한 영상처리기술을 활용하여 금속 표면 검사 공정의 육안 검사를 대체하기 위한 장치개발이 요구된다. 본 논문에서는 압연 공정의 금속 판재 표면의 결함을 탐지하기 위하여 이동평균 영상처리기법을 적용한 정밀측정시스템을 제안하였다. 금속 판재의 압연 처리 과정에서 발생하는 표면 패턴이 결함으로 판정될 수 있다. 이러한 경우에 대하여 제안한 시스템을 적용하여 금속 판재 표면의 실제 결함의 수를 측정할 수 있었고 FFT방법에 의한 측정치에 비하여 양호한 결과를 얻었다. It has been highly required to develope an automatic metal surface inspection system, specifically using image processing techniques, which can replace the visual inspection method in the steel industry. In this paper, we propose a precisional surface measurement system using the moving average image processing technique. When the surface patterns which are generated in the rolling process of metal plates are recognized as defects, the proposed system can measure the actual number of defects. It has been proved that our system shows better results than the conventional FFT method.

金属表面検査における画像処理技術動向

It is difficult for the application of the image processing technology for the processing metal surface inspection to get the satisfaction of the user by various kinds of fluctuation factors in the processing process. We developed an image filter processing to get resemblance degree distribution of every adjacent local domain that was near to algorithm of the human visual inspection. And this image filter was able to emphasize a defect part to the exclusion of various fluctuation factors. As a result, we are able to promote better automation of various processing metal surface inspection.

Quality inspection of metal surfaces by diffractive optical element-based glossmeter

Abstract Planar, convex and concave metal surfaces were produced by utilizing finishing processes that are exploited in the production of tools for plastics injection molding. A novel glossmeter, the so-called diffractive optical element-based glossmeter (DOG), was used for the inspection of the gloss of the surfaces. The Society of the Plastics Industry (SPI) A1 standard, which has the lowest surface roughness of such standards, served as a reference for the success of the finishing process. The results show that by using DOG we can gain local microscopic and macroscopic information on the gloss and its variation. The DOG is sufficiently sensitive to detect small gloss variation as well as the texture of the surface, e.g. anisotropy in surface marks. Some of the surfaces in this study have a higher surface quality than the A1 surface.

Development of Magneto-Optic Eddy Current Probe and Characterization of its Metal Surface Inspection

The purpose of this paper is to introduce a new eddy-current testing technique for surface defect detection in nonmagnetic metals using magneto-optic crystal and exciting coil (magneto-optic eddy current probe, MOECP). The magneto-optic eddy current probe could measure the slit depth of the 0.1 mm with the signal-to-noise ratio (SNR) at 16.8 and detect the slit interval with high spacial resolution.

Image acquisition techniques for automatic visual inspection of metallic surfaces

This paper provides an overview of three different image acquisition approaches for automatic visual inspection of metallic surfaces. The first method is concerned with gray-level intensity imaging, whereby the most commonly employed lighting techniques are surveyed. Subsequently, two range imaging techniques are introduced which may succeed in contrast to intensity imaging if the reflection property across the intact surface changes. However, range imaging for surface inspection is restricted to surface defects with three-dimensional characteristics, e.g. cavities. One range imaging approach is based on light sectioning in conjunction with fast imaging sensors. The second introduced range imaging technique is photometric stereo.

Automated metal surface inspection through machine vision

This paper proposes a method for automated visual inspection of metal surfaces. Firstly, the modified grey-level co-occurrence matrices of metal images are used to access the information of metal surfaces. Secondly, the difference moment and the entropy of the grey-level co-occurrence matrices are extracted as the features of the metal surfaces. Finally, the features of the inspecting images are then compared with the preset confidence interval to determine whether the inspecting metal is defective or not. Some combinations of relative positions between two positioning pixels and feature descriptors were tested in the experiments to find the best one. The experimental results show that the proposed method can detect the defects effectively and has better correct detection rates than the conventional method.

Metal surface inspection using image processing techniques

The feasibility of applying image processing techniques to metal surface inspection is demonstrated. Two methods for metal surface inspection are described. In the first method, the metal surface reflective power and the metal surface normal are related by a random surface scattering model. The metal surface profile can then be computed from the metal surface normal. The second method applies pattern recognition techniques to classify metal surfaces into classes of different roughness. Methods of feature extraction and classification have been tested experimentally and the performances of different types of classifier have been compared. A two-level tree classifier using nonparametric linear classifiers at each node gives better than 90% correct classification on the testing set.

Three‐dimensional, farfield measurements of a Gaussian profile ultrasonic transducer

An ultrasonic transducer has been constructed using a concentric array of annular metallic electrodes on a thin circular piezoelectric substrate. The transducer is driven through an impedance matching and weighting network to produce an electric field within the substrate that is an optimized piecewise linear approximation to a two‐dimensional Gaussian function [P.S. Zerwekh and R. O. Claus, IEEE 1981 Ultrason. Symp., Chicago, IL]. In this paper, the calculated farfield Gaussian amplitude distribution is presented and compared to experimental data measured using a microprocessor‐controlled automatic scanning system at a distance of 500 wavelengths in a water tank. This data is also compared with similar analytical and experimental results obtained for a circular piston transducer having only one active surface electrode. Applications of the transducer in acousto‐optical modulation systems and in the ultrasonic inspection of metallic surfaces is discussed. [Work partially supported by NSF and NASA.]