Defect detection is a crucial task in screen-printed circuit (SPC) production, where image processing method based on deep learning is often used. This field frequently encounters challenges, such as minute surface defects, a large number of model parameters, and high computational complexity. To address these challenges, a self-made SPC defect data set and an enhanced CAAB-YOLOv8n detection algorithm were developed. A CAD module was integrated into the backbone network to improve the model’s ability to detect bar-shaped features. In addition, the ASF feature fusion and RMT modules were combined to construct the ASF-CR neck structure, which enhances the model’s capability to detect small, localized defects. To expedite inference speed, the DBB-Head reparameterization module was incorporated. Experimental results show that the enhanced algorithm achieves 88.4 $$\%$$ % accuracy, a mAP@50 of 90.2 $$\%$$ % , and a parameter count of just 33.27 million, with a detection speed of 35.2 frames per second. The real-time requirements for SPC defect detection are met by these findings. This work lays a solid theoretical foundation for subsequent defect traceability and the optimization of printing process parameters.

Phased transducer arrays (PTA) can control ultrasonic waves to produce a holographic acoustic field. However, obtaining the phase of the corresponding PTA from a given holographic acoustic field is an inverse propagation problem, which is a mathematically unsolvable nonlinear system. Most of the existing methods use iterative methods, which are complex and time-consuming. To better solve this problem, this paper proposed a novel method based on deep learning to reconstruct the holographic sound field from PTA. For the imbalance and randomness of the focal point distribution in the holographic acoustic field, we constructed a novel neural network structure incorporating attention mechanisms to focus on useful focal point information in the holographic sound field. The results showed that the transducer phase distribution obtained from the neural network fully supports the PTA to generate the corresponding holographic sound field, and the simulated holographic sound field can be reconstructed with high efficiency and quality. The method proposed in this paper has the advantage of real-time performance that is difficult to achieve by traditional iterative methods and has the advantage of higher accuracy compared with the novel AcousNet methods.

The method of physically recording acoustic holograms is time-consuming and inefficient, hindering the ability to rapidly reconstruct acoustic holograms. This text aims to propose a deep learning strategy based on an improved IU-Net to rapidly reconstruct acoustic holograms with phase information. The angular spectrum method is used to simulate the phase information distribution of different sound sources, positions, and initial angles, thereby quickly generating the required data samples, especially for complex or dynamic image scenarios. The framework employs a custom learning rate adjustment mechanism to dynamically adjust the learning rate in real-time, adapt to the training process, and support a stepwise learning strategy. Experimental results show that the proposed method achieves an accuracy of 90.9% in acoustic hologram reconstruction, with a Dice coefficient of 0.928, improving reconstruction speed by over 30% compared to traditional methods, ensuring high-quality and rapid reconstruction of acoustic holograms.

The ability to reconstruct acoustic field is essential for applications such as particle manipulation, medical therapy, and ultrasonic imaging. Conventional acoustic field reconstruction algorithms are limited in both speed and accuracy when generating acoustic holograms, which further constrains the practical applications of acoustic holography. In this study, we propose an acoustic field reconstruction method that achieves both speed and accuracy by leveraging an acoustic model tailored for diffraction. A unit transducer is integrated with an acoustic lens, utilizing lens thickness and acoustic velocity variations to produce accurate phase maps at specific frequencies. To reduce the dependence on labeled datasets, we propose a deep learning framework with an autoencoder structure for unsupervised training without relying on any a priori information. Additionally, a novel loss function is designed for the framework to realize the energy efficiency constraint. The proposed method is compared with traditional Angular Spectrum Method (ASM) and Automatic Differentiation (Diff-PAT). Experimental results show that the PSNR of this method reaches 21.59, which is 24.1% higher than that of ASM and 19.2% higher than that of Diff-PAT. After network training, the proposed method generates phase holograms hundreds of times faster than ASM and Diff-PAT. Under the same network architecture, our newly designed loss function improves energy efficiency to 2.1 times that of the original method.

Compared with a normal flame detector system,IUR76-Ⅰ/Ⅱ test systems optimize the system architecture and stability based on related national standards. It is proved that the new test system is more stabile than the normal test system. The stabilities of the new test system and the system based on national standards are calculated and compared. The suggestions according to the related national standards are also mentioned.