Surface Casting Defects Inspection Using Vision System and Neural Network Techniques

Świłło, S.; Perzyk, M.

doi:10.2478/afe-2013-0091

Cited by 19 publications

(13 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This comparative metric is calculated as the maximum phase correlation according to (7), where F stands for Fourier transform of images a, b, F -1 is the inverse Fourier transform, and F* is the complex conjugate image [20].…”

Section: Poc Phase-only Correlationmentioning

confidence: 99%

“…Automated visual measuring is known to be an imageprocessing method that has been widely applied over the years in the production line for quality control purposes [4], [5]. This focuses on mechanical parts, vehicles, casting production and even the garment industry [6], [7]. Inspection tasks, however, are time consuming and have been usually and mostly carried out through human intervention, i.e.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Measuring and evaluating the differences of compared images for a correct car silhouette categorization using integral transforms

Pivarčiová

Sobrino

Nikitin

et al. 2018

Measurement Science Review

View full text Add to dashboard Cite

The present paper focuses on the analysis of the possibilities of using integral transforms for measuring and evaluating the differences of compared images (car silhouettes) with the purpose of a correct car body categorization. Approaches such as the light intensities frequency change, the application of discrete integral transforms without the use of further supplementary information enabling automated data processing using the Fourier-Mellin transforms are used within this work. The calculation of the several metrics was verified through different combinations that implied using and not using the Hamming window and a low-pass filter. The paper introduced a method for measuring and evaluating the differences in the compared images (car silhouettes). The proposed method relies on the fact that the integral transforms have their own transformants in the case of translation, scaling and rotation, in the frequency area. Besides, the Fourier-Mellin transform was to offer image transformation that is resistant to the translation, rotation and scale.

show abstract

Section: Poc Phase-only Correlationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Measuring and evaluating the differences of compared images for a correct car silhouette categorization using integral transforms

Pivarčiová

Sobrino

Nikitin

et al. 2018

Measurement Science Review

View full text Add to dashboard Cite

show abstract

“…With the rapid expansion of network applications, computer vision technology has been successfully applied to the quality inspection of industrial production [1][2][3][4][5][6], including glass products [1], fabrics [2,3], steel surfaces [4], bearing rollers [5], and casting surfaces [6]. The inspection of these mentioned examples needs a matching algorithm to extract image features based on the actual defect situation.…”

Section: Introductionmentioning

confidence: 99%

“…For the inspection of surface aluminum, a vision based approach and neural network techniques in surface defects inspection and categorization are proposed. The new vision inspection system, image processing algorithm, and learning system based on artificial neural networks (ANNs) were successfully implemented to inspect surface aluminum die casting defects [6].…”

Section: Introductionmentioning

confidence: 99%

An Intelligent Vision System for Detecting Defects in Micro-Armatures for Smartphones

et al. 2019

View full text Add to dashboard Cite

Automatic vision inspection technology shows a high potential for quality inspection, and has drawn great interest in micro-armature manufacturing. Given that the inspection process is highly influenced by the lack of real standardization and efficiency performed with the human eye, thus, it is necessary to develop an automatic defect detection process. In this work, an elaborated vision system for the defect inspection of micro-armatures used in smartphones was developed. It consists of two parts, the front-end module and the deep convolution neural networks (DCNNs) module, which are responsible for different areas. The front-end module runs first and the DCNNs module will not run if the output of the front-end module is negative. To verify the application of this system, an apparatus consisting of an objective table, control panel, and a camera connected to a Personal Computer (PC) was used to simulate an industrial position of production. The results indicate that the developed vision system is capable of defect detection of micro-armatures.Recently, deep convolutional neural networks (DCNNs) have been proved to be important methods in visual detection. However, some classical methods should still be considered. Literature [6] mentioned above is an example. For instance, in order to get better parameters, a suitable smart manufacturing strategy for real industrial conditions was proposed. The results of this dataset showed that the Adaboost ensembles provided the highest accuracy and were more easily optimized than ANNs [7]. Obviously, for now, there are some limitations with using classical methods alone or using DCNNs directly. For classical methods, for example, they usually come with high complexity of programming and less tolerance to data variability. With regard to the DCNNs, a large number of samples are needed for network training and the iterative optimization of parameters is partly a black-box operation [8].In the industrial applications, part defects in the samples can be easily identified with a picture by the classical computer vision method. Therefore, we do not need to feed the whole picture into the network for discrimination, which can reduce the number of features to be identified. As a result, the network is easier to converge. Therefore, by combining the classical image recognition method and DCNNs, the landing speed of deep learning technology in industrial applications can be effectively improved.In this work, an intelligent detection system that combined the classical computer vision method and DCNNs was designed to automatically detect the quality defects of micro-motor armatures. Firstly, the quality, excluding the region of copper wire crossing (ROC), is decided based on the classical computer vision technology. If the first result is positive, the ROC will be extracted and sent to the DCNNs for identification. If the result is still positive, the image is a defect-free sample, otherwise the whole image is labeled defective. In the experiments, this system works very fast and presen...

show abstract

“…are basic parameters in fuzzy logic algorithm, in which new parameters like radius ratio, axis ratio and approximate area were calculated to characterize the shape feature of defect area. Świłło [2][3] et al developed a surface defect inspecting machine for die castings with image processing algorithms based on modified Laplacian of Gaussian edge detection method to recognize defects with different shapes and sizes. Most of the relevant works were carried out by analyzing the optical images taken from the object surfaces and tried to classify surface defects by calculating shape features like area, perimeter, length, etc.…”

mentioning

confidence: 99%

Quantitative evaluation of die casting surface defect severity by analyzing surface height

Hirata

Anzai

2017

China Foundry

View full text Add to dashboard Cite

Male, Ph.D. candidate in Tohoku university. His research interests mainly focus on the quantitative analysis for microstructure of die castings and compounds using mathematical morphology. V arious types of surface defects of die casting products may occur due to the complex situation inside the die. The evaluation and classification of surface defects on die castings have long been an important and challenging issue. Traditionally, visual inspection depending on technicians' individual experience has been used in factories to examine whether the surface defects are acceptable or not. However, this calls for high requirement on the technicians' skill. Even professional technicians will find it difficult to balance the conflict between efficiency and accuracy, since the work is so subjective, tedious and monotonous.Image processing technology provides a viable alternative to automatically detect the defects and evaluate the defect severity. Wong [1] et al. adopted fuzzy logic algorithm to surface images after image Abstract: It is necessary in factories to assess the severity of the surface defects of castings, as a slight surface defect will be taken as qualified when it brings no bad effect or it can be removed by the subsequent processing. In practical production, professional technicians visually inspect the surface defect severity according to their individual experience. Therefore, it is difficult for them to maintain the same standard and accuracy in the subjective, tedious and labor-intensive work. Recently, image processing techniques based on optical images have been applied to achieve better accuracy and high efficiency. Unfortunately, optical images cannot directly quantify surface depth, which works as a crucial factor in the practical assessment of surface defect severity. The surface roughness evaluation algorithm, which takes into account of both area and depth information of the assessed surface, was applied to directly characterize surface defect severity based on surface asperity rather than optical image. The results using standard casting pieces show that surface defect severity has no apparent dependence on surface roughness. However, the subsequent results show that the root-mean-squared-deviation (RMSD) of surface gradient of flow line defects positively correlates with the increase of defect severity. The other types of defect do not present such tendency. Thus, practical workpieces with flow line defects on the surface were used to verify the universality of this tendency. The results show that surface roughness of an unqualified workpiece is larger than that of a qualified workpiece after surface slope adjustment, but presents no obvious coincidence before the adjustment. In contrast, the RMSD of an unqualified workpiece, no matter before or after the adjustment, is larger than that of a qualified one.

show abstract

Surface Casting Defects Inspection Using Vision System and Neural Network Techniques

Cited by 19 publications

References 5 publications

Measuring and evaluating the differences of compared images for a correct car silhouette categorization using integral transforms

Measuring and evaluating the differences of compared images for a correct car silhouette categorization using integral transforms

An Intelligent Vision System for Detecting Defects in Micro-Armatures for Smartphones

Quantitative evaluation of die casting surface defect severity by analyzing surface height

Contact Info

Product

Resources

About