Due to the frequent occurrence of missed detections, false alarms, and the low accuracy of surface defect detection in industrial components, surface defect detection has always been a challenge in the industrial field. In this study, we propose a novel network structure based on YOLOv5s that achieves improved defect detection accuracy by improving the YOLOv5s network structure. First, a CoTNet transformer module is incorporated in the feature extraction process of the backbone network, replacing the original C3 network module. A global attention mechanism (GAM) is integrated along with the C3 module into the neck to enhance the feature learning ability of the model so that it focuses more on features. Last, the adaptive spatial feature fusion algorithm is used for prediction to increase the forecasting accuracy of the model. Finally, the improved algorithm is evaluated on a dataset containing NEU-DET steel surface defects. The experimental results show that the improved algorithm in this paper outperforms other current state-of-the-art algorithms in terms of on this dataset detection accuracy. In addition, we conduct additional experiments on the PKU-Market-PCB dataset, confirming that the proposed method yields improved detection accuracy.
Stable rheological properties of ceramic ink are a key requirement for inkjet printing (IJP), which should be satisfied in terms of the Reynolds and Weber numbers. In this paper, the reverse microemulsion was introduced for the synthesis of monodispersed nanosized ceramic powders, and the average size was less than 100 nm. A comparison of two different dispersants, i.e., polyacrylic ammonium (PAANH4) and polyacrylic aid (PAA), revealed that the former exerted a good dispersion effect on the ceramic ink. The sedimentation ratio, zeta potential, surface tension, viscosity, and density of the inks were measured, and the Reynolds and Weber numbers, as well as Z value, were calculated. A stable, homogeneous, and high solid loading (20 wt%) ceramic ink could be achieved after aging for a period of 72 h. Finally, the ceramic inks showed the desired printable property in the inkjet printing process. Combining inkjet printing technology with a sintering process, Ni-Mn-O films have the potential to monitor temperature and humidity parameters for intelligent wearable devices.
The use of the thermostatic metal bath is becoming more and more extensive and the requirements for its precision and reliability are also increasing. To meet the needs of the metal bath calibration, a 12-channel thermostatic metal bath temperature field calibration system based on a four-wire PT100 has been designed. The system design includes a front-end temperature measurement component, which consists of a four-wire PT100 and a thermostatic block, and a signal processing component, which consists of a bidirectional constant current source excitation unit, a signal conditioning unit and a high-precision acquisition unit. The STM32f407 is used as the main control chip, and the analog channel selector is used for 12-channel selection. The constant current source is used for signal excitation, the proportional method is used to measure the PT100 resistance, and an acquisition circuit with a high-precision 32-bit ADS1263 analog-to-digital converter is used to amplify, filter and convert the analog signal. After piecewise linear fitting and system calibration, the temperature measurement accuracy can reach 0.4 mK, which meets the calibration requirements of the thermostatic metal bath.