Search

Over the last few years, advanced deep learning-based computer vision algorithms are revolutionizing the manufacturing field. Thus, several industry-related hard problems can be solved by training these algorithms, including flaw detection in various materials. Therefore, identifying steel surface defects is considered one of the most important tasks in the steel industry. In this paper, we propose a deep learning-based model to classify six of the most common steel strip surface defects using the NEU-CLS dataset. We investigate the effectiveness of two state-of-the-art CNN architectures (MobileNet-V2 and Xception) combined with the transfer learning approach. The proposed approach uses an ensemble of two pre-trained state-of-the-art Convolutional Neural Networks, which are MobileNet-V2 and Xception. To perform a comparative analysis of the proposed architectures, several evaluation metrics are adopted, including loss, accuracy, precision, recall, F1-score, and execution time. The experimental results show that the proposed deep ensemble learning approach provides higher performance achieving an accuracy of 99.72% compared to MobileNet-V2 (98.61%) and Xception (99.17%) while preserving fast execution time and small models’ size.

Often, surface mechanical treatment (SMT) or heat (quenching, tempering) is used to improve the surface condition and mechanical characteristics such as impact resistance and tensile strength. Hence the objective of this experimental work, where ball burnishing (BB) as well as quenching and tempering were applied on S355 JR steel specimens, is to evaluate the surface hardness Hv, the rupture strength Rm, the energy absorbed W during the impact test, and the work-hardened thickness e after the burnishing operation. Factorial designs were used for the test organization and mathematical models were established for the prediction of Hv, Rm, W, and “e” in function on treatment parameters considered: number of tool passes (i) and the pressure force (Py). The results show that the surface hardness increases by 30.46%. The high levels of Py and i were allowed to improve Rm by 30.8% as well as an increase in tenacity of only 3.6%. Increasing the force to 20kgf promotes mixed rupture under the effect of impact to shock. The quenching and tempering improve the Rupture strength Rm by 183% and 119%, respectively, while the effect of burnishing was limited to a rate of increase of this property of 28% compared to machining.

The shaped steel strip, in the hot rolling process, may exhibit some surface flaws. Their origin could be the internal discontinuities in the input product or the thermomechanical transformation of the material, during the shaping process. Such defects are of a random occurrence and may lead to costly rework operations or to a downgrading of the final product. So, they should be detected and identified as soon as possible, to allow a timely decision-making. For such a quality monitoring, the used vision systems are mainly based on an image description and a reliable classification. In this paper, we explore pre-defined image filters and work on a procedure to extract a discriminant image feature, while realizing the best trade-off between the improved recognition rate of the surface defects and the computing time. The proposed method is a multiresolution approach, based on the Binarized Statistical Image Features method, employed to date in biometrics. The filters, pre-learnt from natural images, are applied to steel defect images as a new surface structure indicator. They provide a quite discriminating image description. A relevant data reduction is used together with a classifier to allow an efficient recognition rate of the defective hot rolled products.

In steel-making processes, different methods are used for online surface product monitoring. Such a control has become a necessity to avoid additional costs resulting from the poor quality of the final product. With the reported performance that varies from one application to another, all the applied methods have to meet a minimum of criteria as accuracy and speed. This effectiveness is assured thanks to a relevant image description and efficient defect classification algorithms. The Dual Cross Pattern technique, successfully applied in face recognition, is a concept that relies on coding pixels to provide such a discriminating description of the image. Its principle can perfectly be used in industrial vision applications for surface defect recognition. In this study, the relevance of this method of describing defect images is evaluated, and improvements are proposed to increase its efficiency. The experimental study shows that the pixel coding that considers the variations of the intensity in several directions and captures the information from more than one pixel-neighborhood level makes it possible to better detect the variability in the defect image and helps to increase the defect recognition rate. The experiments are carried out with the use of the published Northeastern University (NEU) database for the comparison and with a new constructed database to better show the improvements brought by the proposed approach.

Over the last decade, palmprint recognition has been studied for many problems and applications. It has become one of the most well-known biometric recognition system. Its success is due to the rich features that can be extracted and exploited from the palmprint images captured by contact or contactless device. This paper presents a new representation based on textural structure of human palms for touchless palmprint identification. This representation method is called Patch Cross Pattern (PCP), which relies mainly on cross pattern encoder and the non-overlapping decomposition method. The feature vector is built using Cross Pattern (CP) encoder to capture the textural structure of palmprint image. Then, the non-overlapping decomposition on both directions is applied. Next, the feature vector representation of each palmprint image is constructed by concatenating all normalized histograms calculated at each patch. In addition, the reduced version of the PCP called R-PCP is obtained using whitened linear discriminant analysis. Finally, a K-nearest neighbor classifier is used for palmprint identification. The proposed system is successfully applied to IIT Delhi and CASIA touchless databases. Results show that, the proposed representation provides a significant performance improvement compared to the recent state-of-the-art in terms of accuracy.

The evaluation of flat steel surface quality is mainly concerned with detecting and identifying product surface defects. Although the variety of the implemented techniques, this type of control still presents a challenge. In this paper, we assess the Dual Cross Pattern technique, as a feature descriptor, that should be quite discriminative, to ease the steel surface defect classification. The histograms extracted from the captured DCP features are concatenated to represent the global image feature vector. The procedure parameters, as the DCP circle radius, the number of the training images and their choice, are considered to show their impact on the results. The experiment conducted on the NEU published defect database shows that, compared to the other used techniques, the proposed approach reveals not only interesting recognition rates but presents advantages in time coast too.

This paper is concerned with a method for uncertainty evaluation of mechanical properties in metal testing. This method uses a combined approach based on Monte Carlo simulation and Markov Chain (MCMC) as a computing procedure of different uncertainties of mechanical and metallurgical parameters such as stress, and elongation. The MCMC is a stochastic method that computes the statistical properties of the considered states such as the probability distribution function (PDF) according to the initial state and the target distribution using Metropolis-Hasting (MH) algorithm. Conventional approach is based on the Guide of Uncertainty Measurement (GUM), the uncertainty budget is established for the stress and elongation parameters respectively. A comparative study between the conventional procedure and the proposed method is given. This kind of approaches is applied for constructing an accurate computing procedure of uncertainty measurement of mechanical and metallurgical parameters.

Vision techniques have become, since many years, a relevant means for mastering product quality parameters, as dimensions, especially in on line control and in a hard working environment. This development has been reached thanks to technologic progress in equipements, sensors and data processing power. Indeed, in dimensionnal control, industrial vision allows the detection of specific points of the object, to compute the sought dimension. This latter is, in a cartesian coordinates system, the two points Euclidean distance. However, depending of the application type and in spite of their performance, used cameras present, some constraints as parallaxe error. It is due to the natural shift, at the installation, of the cameras optical axis regarding target object; or to the dynamic state of the object. This error may affect significantly the measure precision. Stereoscopic vision is a solution that allows minimizing the above error and presenting the compromise between precision, installation space and costs. Two cameras are used to observe, each one, the whole object to be measured. In the proposed practical approach, the application is highly depending of a calibrating step, that consists in creating a data base composed of a maximum image points related to the target plan and their geometric positions in a considered coordinates system. When executed, the algorithm detects, thanks to the two cameras acquired information, pixels corresponding to object edges, find their correspondance in the created data base, computes their coordinates (xi,yi) in the defined coordinates system to, finally, computes the sought dimension. 