On the development of intelligent optical inspections

Richter, Johannes; Streitferdt, Detlef; Rozova, E. Yu.

doi:10.1109/ccwc.2017.7868455

Cited by 33 publications

(15 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These methods can be used to monitor the operating parameters of the cable, such as its frequency and transmittance properties, as well as the consumption of insulating materials. To implement visual-optical inspection methods, it is possible not only to use direct visual observation of cable quality by staff (manual optical control), but also to use computer vision-based intelligent control [10][11][12]. The disadvantage of the visual and optical fault location methods is that their use in implementing full control of parameters of electrically conductive electrical cable elements is significantly limited.…”

Section: Analysis Of Well-known Electrical Cable Fault Location Methodsmentioning

confidence: 99%

Nondestructive Testing Method Based on the Spin Polarization Effect Template

2020

View full text Add to dashboard Cite

This paper proposes a novel nondestructive testing method for electrically conductive objects. It is based on applying a fundamentally new physical effect which has not been used before in nondestructive testing systems. The method utilizes spin polarization phenomena that occur when free electrons of conductive materials are exposed to an alternating electric field. Researchers experimented with wire and cable conductors. A tested cable was moving longitudinally and exposed to an alternating electric field, which excited a waveform process, i.e. the polarization of spin magnetic moments of free electrons. An induction sensor registered this process and generated a control signal: induction EMF. Its parameters were compared against that of a reference signal obtained in advance in the same way utilizing spin polarization phenomena. Parametric deviations were then used to detect and classify conductor defects. The paper describes how to generate a reference signal appropriate for the objectives of testing. The developed method enables nondestructive testing of objects made of any conductive para-and diamagnetic materials, while it accuracy and reliability are not affected by the magnitude or evenness of the object movement speed, nor by vibrations or transverse oscillations against the physical field source or the induction sensor.

show abstract

Section: Analysis Of Well-known Electrical Cable Fault Location Methodsmentioning

confidence: 99%

Nondestructive Testing Method Based on the Spin Polarization Effect Template

2020

View full text Add to dashboard Cite

show abstract

“…There have been a wide range of studies on inspecting defects in PCB from using traditional machine learning methods such as random forest [30] to using more modern deep learning approach, utilizing multi-layer perceptron neural network and convolutional neural network [31], [32]. Another neural network based study incorporates fuzzy rule-based method to correct any possible misclassification made by the neural network module [33].…”

Section: Smt Inspectionmentioning

confidence: 99%

MarsNet: Multi-Label Classification Network for Images of Various Sizes

et al. 2020

View full text Add to dashboard Cite

Since the Convolutional Neural Network (CNN) has surfaced and fascinated the world, many researchers have exploited CNN for image classification, object detection, semantic segmentation, etc. However, the conventional CNNs have a pyramidal structure and were designed to process images which have the same size. Although some CNNs can accept images of various sizes, performance is degraded for images smaller than the size of images used for training. In this paper, we propose MarsNet, a CNN based end-to-end network for multi-label classification with an ability to accept various size inputs. In order to allow the network to accept such images, dilated residual network (DRN) is modified to get higher resolution feature maps, and horizontal vertical pooling (HVP) is newly designed to efficiently aggregate positional information from the feature maps. Furthermore, multi-label scoring module and threshold estimation module are employed to serve the purpose of multi-label classification. We verify the effectiveness of the proposed network through two distinctive experiments. We first verify our model by inspecting and classifying multiple types of defects occurred in PCB screen printer using solder paste inspection (SPI) datasets. Secondly, we verify our network using VOC 2007 dataset. Our network is pioneering in that no research has attempted to accomplish multi-label classification for defects in addition to being able to take input images of various sizes in SPI field. INDEX TERMS Convolutional neural networks, images of various sizes, multi-label classification, printed circuit board, solder paste inspection.

show abstract

“…Supported by the advent of powerful computational devices, e.g., graphical processing units (GPUs), deep learning-based techniques have become essential tools in fault detection and fault diagnosis research fields. Their superior performance in applications related to classification and object detection tasks has also supported their popularization [4].…”

Section: Introductionmentioning

confidence: 99%

Using a Support Vector Machine Based Decision Stage to Improve the Fault Diagnosis on Gearboxes

Monteiro

Cerrada

Cabrera

et al. 2019

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Gearboxes are mechanical devices that play an essential role in several applications, e.g., the transmission of automotive vehicles. Their malfunctioning may result in economic losses and accidents, among others. The rise of powerful graphical processing units spreads the use of deep learning-based solutions to many problems, which includes the fault diagnosis on gearboxes. Those solutions usually require a significant amount of data, high computational power, and a long training process. The training of deep learning-based systems may not be feasible when GPUs are not available. This paper proposes a solution to reduce the training time of deep learning-based fault diagnosis systems without compromising their accuracy. The solution is based on the use of a decision stage to interpret all the probability outputs of a classifier whose output layer has the softmax activation function. Two classification algorithms were applied to perform the decision. We have reduced the training time by almost 80% without compromising the average accuracy of the fault diagnosis system.

show abstract

On the development of intelligent optical inspections

Cited by 33 publications

References 0 publications

Nondestructive Testing Method Based on the Spin Polarization Effect Template

Nondestructive Testing Method Based on the Spin Polarization Effect Template

MarsNet: Multi-Label Classification Network for Images of Various Sizes

Using a Support Vector Machine Based Decision Stage to Improve the Fault Diagnosis on Gearboxes

Contact Info

Product

Resources

About