Transfer learning-based and originally-designed CNNs for robotic pick and place operation

Herein, we studied the increasing tendency of photoacoustic (PA) conversion efficiency of the Au/polydimethylsiloxane (PDMS) composite. The thickness of the Au layer was optimized by modeling the PA process based on the Drude-Lorentz model and finite element analysis method, and corresponding results were verified. The results showed that the optimal Au thickness of the Au/PDMS composite was 35 nm. Finally, the Au/PDMS composites were coated onto the surface of aluminum alloys, which improved the thermoelastic laser ultrasonic (LU) signals to near 100 times. Besides, the defect mapping was performed by thermoelastic LU signals with Au/PDMS coating and ablation LU signals without coating; the Pearson correlation coefficient was higher than 0.95. The application in the defect detection in metal could provide guides for nondestructive detection on metals by laser ultrasound.

In this paper, a simple and efficient edge detection technique is proposed. It is based on the properties of the hyper smoothing function and the fundamentals of modified local binary pattern and hence, called as HY-LBP. The main advantage of the proposed approach is that the relationship between surrounding pixels and centric one is calculated effectively and extract the surrounding information discriminatively. The counting scheme that counts the number of image points whose pixel values are greater or equal than that of the central pixel help to reduce the noise and blurring effects. Thus, HY-LBP can effectively deal with the noises, blurring, and contrast variation. The effectuality and feasibility of the proposed approach is demonstrated on various synthetic and real time images. Experimental results have shown the effectiveness as well as the better performance of the proposed approach for edge detection as compared to the other existing methods. The presented method can be used in metal sheet defect detection applications.

Defect detection in power systems is crucial for ensuring the safety, stability, and reliable operation of these systems. However, traditional image recognition algorithms face challenges such as low accuracy, slow speed, and poor generalization when dealing with large-scale multi-object detection tasks. This paper proposes an enhanced YOLOv8 algorithm for large-scale multi-object detection and applies it to defect detection in power systems. The proposed method introduces attention mechanisms, multi-scale feature fusion, and adaptive anchor box generation techniques based on the YOLOv8 algorithm, thereby improving the model’s detection accuracy and robustness. Experiments were conducted on a publicly available power system image datasets and compared with algorithms such as YOLOv8, YOLOv5, and Faster R-CNN. The experimental results demonstrate that the proposed method has a significant advantage in large-scale multi-object detection, achieving a mean average precision (mAP) of 90.2.

摘要： 针对小径管缺陷的检测，传统检测方法基本为Bobbin探头，但这种方法具有不能检出缺陷的环向信息的缺点。基于涡流检测的研究，为达到检出小径管缺陷的多重信息的目的，提出并开发一种新型的环向偏心Bobbin探头。该探头兼具轴向扫描和偏心环向扫描的功能，能够对小径管缺陷的多重信息进行有效检测。具体研究内容基于涡流检测的方法，设计出适用于小径管的可偏心、可环向旋转的Bobbin探头，并建立模型进行模拟计算；通过整合获得的模拟数据，验证新型探头对于小径管缺陷的轴向、环向扫描相比于传统探头的优越性；制作探头并搭建试验系统，对所开发探头检测小径管缺陷的功能进行验证，进一步确立了该新型探头的优越性和高效性。

Reliable leather defect detection is essential for industrial quality control, yet many deep learning solutions demand substantial computational resources and extensive training. This work addresses automatic defect classification on the public <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Leather Defect Classification</i> dataset (3,600 images across six classes), seeking high accuracy with low computational cost. We propose <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Cosine–Histogram Association</i> (CHA), an ultra-lightweight approach based on ℓ<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>-normalized RGB color histograms and cosine similarity. In direct comparison on the same dataset, CHA (using 17 × 45 × 36 histogram bins) achieved 97.08% test accuracy, surpassing the best convolutional model trained from scratch in this study (92.4%) and prior reports for widely adopted architectures (AlexNet, VGG16, ResNet-50, Inception-v3), which achieved 80.0–87.6% test accuracy on the same dataset. In addition, CHA processed 2,520 training images in 0.79 s and classified 720 test images in 0.40 s, totaling 1.19 s for 3,240 images—up to 177× faster than CNN training and 4.5× faster than CNN evaluation. These results indicate that CHA can deliver strong discriminative performance while substantially reducing computational requirements, supporting its use for automated visual inspection in resource-constrained industrial settings.

According to the characteristics of composite materials and its possible application working environment, to make sure of using the compound material safely, through a method to test the defect of compound material is quite important. With the development of electronic technology, ultrasonic phased array is gradually used in industry nondestructive testing. This article analyses the advantages of ultrasonic phased array, choosing the ultrasonic phased array technology as the nondestructive testing method in test Compound material defect, to receive and transmit ultrasonic phased array in the process of the calculation of delay time of the array element required to control deflection and focusing of acoustic beam, and to determine the appropriate parameters of phased array probe, in order to achieve the best effect of detection. Finally, to test the model composite, and to validate the advantages of ultrasonic phased array technology in the detection of new composite materials, and has the good prospects for development.

In addition to causing damage to vehicles, road defects are one of the main causes of vehicle accidents which lead to loss of human lives. Many methods of detecting defects have been introduced over the years to reduce the consequences of these defects. One of these methods is image processing. Use of image and video processing has many applications in medicine, science, agriculture, and defect detection in structures. It has been used for defect detection on roads because timely detection and analysis of defect is very important for road serviceability and safety of the people. Detection of a defect by image processing broadly follows some of the basic steps which include feature extraction, edge detection, morphological operators, and training of data. Different approaches are used for various kinds of defect detection and analysis which have replaced the manual inspection method of roads saving time and resources. This chapter discusses the basic steps involved in defect detection using image processing along with existing systems that use machine learning and artificial intelligence for the detection of defects from a distance. To write this chapter, papers on the topic of image and computer vision-based defect detection systems have been consulted.

Glycoproteins pose significant challenges for specific recognition due to their structural complexity and microheterogeneity. Molecularly imprinted polymers (MIPs) have emerged as promising synthetic receptors, offering high stability and tailorable recognition sites. This review specifically highlights and systematically evaluates several cutting-edge design strategies tailored for glycoproteins, including oriented surface imprinting for site-accessible recognition, epitope imprinting for enhanced specificity, and post-imprinting modification for tailored functionality. The fundamental principles, technical advantages, and applications in glycoprotein detection and separation are thoroughly discussed, with a particular emphasis on a comparative analysis to guide strategy selection and how they collectively address the persistent challenges of traditional imprinting. Future perspectives highlight stimuli-responsive systems, multimodal recognition, and computational design to advance MIPs as indispensable tools in proteomics and personalized medicine. The synergistic integration of these advanced strategies within sustainable and standardized MIP systems is particularly promising for fabricating next-generation synthetic receptors with enhanced recognition capabilities.

Triple-mode biosensing strategies: A critical review of design principles, performance metrics, and applications for ultrasensitive detection

Circularly polarized light (CPL) with the end of optical vector traveling along circumferential trajectory, shows left- and right-handedness, which transmits chiral information to materials via complicated CPL-matter interactions. Materials with circular dichroism respond to CPL illumination selectively with differential outputs that can be used to design novel photodetectors. Racemic or achiral compounds under CPL go through photodestruction, photoresolution and asymmetric synthesis pathways to generate enantiomeric bias and optical activity. By this strategy, helical polymers and chiral inorganic plasmonic nanostructures are synthesized directly, and their intramolecular folding and subsequent self-assembly are photo-modulable as well. In the aggregated state of self-assembly and liquid crystal phase, helical sense of the dynamic molecular packing is sensitive to enantiomeric bias brought by CPL, enabling the chiral amplification to supramolecular scale. In this review, we aim to systematically summarize and discuss the application-guided design strategies of CPL-responsive materials. Asymmetric synthesis, resolution and property-modulation of small organic compounds, polymers, inorganic nanoparticles, supramolecular assemblies and liquid crystals are highlighted based on the important developments during the last decades. Besides, applications of light-matter interactions including CPL detection and biomedical applications are also referred. This article is protected by copyright. All rights reserved.

Abstract: Isoniazid-based fluorogenic and chromogenic chemosensors have gained considerable attention for their rapid, selective, and visually detectable responses toward metal ions and anions. This review presents a comprehensive analysis of their structural design, sensing mechanisms, and practical applications. The electron- rich framework of isoniazid facilitates hydrogen bonding and coordination complex formation, making it an excellent scaffold for chemosensor development. These systems have been successfully applied in environmental monitoring, medical diagnostics, and cellular imaging, demonstrating high sensitivity and versatility. Critical parameters such as detection limits, analyte selectivity, solvent effects, and sensing modes are discussed to guide future research and development in this growing field of chemical sensing.

As a crucial component of the photovoltaic (PV) power generation system, effective defect detection of PV cells can help to ensure the power generation efficiency and the system safety. The photo-thermal-radiation infrared thermography experimental system has been built, and the infrared images of normal and defective PV cells were collected by an IR camera. An effective convolutional neural network (CNN) based model with the residual architecture is designed here to detect and classify the defects. An offline date augmentation method is performed to overcome the insufficient image dataset in the step of dataset establishment, so as to improve the defects classification capability of the designed CNN model. The defects classification accuracy of the proposed model is obtained as 96.96% with 3-class. In order to verify the defect classification ability of the designed CNN model, other four models were adopted for comparative analysis. The results indicate that the proposed CNN model has advantages in defect classification of PV cells. The completed work can provide meaningful exploration in the industrial application of defect detection of PV cells.

We present an automated application for defect detection and classification from ZEISS multibeam scanning electron microscope (MultiSEM®) images, based on machine learning (ML) technology. We acquire MultiSEM images of a semiconductor wafer suited for process window characterization at the imec iN5 logic node and use a dedicated application to train ML models for defect detection and classification. We show the user flow for training and execution, and the resulting capture and nuisance rates. Due to straightforward parallelization, the application is designed for the large amounts of data generated rapidly by the MultiSEM.

Thin films are widely used in many fields such as electronics, optics or materials science. For example, they find applications in thermal or mechanical sensors design. They are also very useful as protective or reinforcement layers for many structures. However, some coating defects such as thickness variations, microfissuring or poor adhesion are common problems. Therefore, nondestructive testing of these structures using acoustic waves generated and detected by lasers represents a major interest. Indeed, in comparison with conventional methods based on the use of piezoelectric transducers, laser ultrasonics leads to non-contact investigations with a large bandwidth. Usually, bulk acoustic waves are used and a pulse-echo technique is considered that needs high frequencies and implies local measurements. In order to avoid this limitation, we propose to use surface acoustic waves in a frequency range up to 45 MHz. The samples consist of a micrometric gold layer deposited on silicon substrates. In a first part, using dispersion analysis, theoretical and experimental results clearly reveal that the first Rayleigh mode allows the detection of film thickness variations and open cracks. In a second part, a localized adhesion defect is introduced in a similar sample. The effects of such a flaw on the Rayleigh modes dispersion curves are theoretically described. Finally, we experimentally show that the first Rayleigh mode allows the defect detection only under specific conditions.

• Constructed the largest multi-category dataset for tunnel lining defect detection following rigorous data collection, analysis, and processing. • Proposed an improved KNet-based segmentation model that is validated to outperform multiple composite algorithms under a consistent transfer learning strategy. • Analyzed the model's decision-making process and performance factors in depth via progressive visual interpretability and multi-index joint analysis. • Developed a desktop application for the Central Yunnan Water Diversion Project to facilitate the efficient detection and identification of lining defects in experimental sections. During tunnel construction and operation, the timely detection and treatment of lining defects are critical to ensuring the safety and stability of tunnel engineering. To achieve fast and accurate detection of multi-category tunnel lining defects from image data, this paper proposes an improved model based on the KNet segmentation algorithm, integrated with the Swin Transformer and UPerNet models. First, a large-scale multi-category image dataset of tunnel lining defects is constructed by collecting and organizing data from relevant literature and public datasets. Second, the pre-trained KNet-enhanced model is trained on this dataset using transfer learning to improve detection accuracy and generalization performance. The proposed KNet-UPerNet-SwinL model achieves an average improvement of approximately 16.94% in mean Intersection over Union (mIoU) compared to other classical semantic segmentation methods, demonstrating superior segmentation accuracy and stability. In addition, an in-depth discussion of the model’s decision-making process and the factors influencing its performance is provided through visually interpretable progressive feature analysis and multi-index joint analysis at the application level. Finally, to facilitate practical application, the model is converted into the Open Neural Network Exchange (ONNX) universal format, and a desktop application for tunnel lining defect detection is developed. The detection and recognition of lining defects are successfully implemented in the experimental sections of the Central Yunnan Water Diversion Project, achieving remarkable segmentation results. In summary, this study provides a comprehensive solution that offers practical technical and tool support for multi-category defect detection in tunnel linings and holds significant importance for the safety and stability of tunnel engineering.