Title: Industrial phased array ultrasonic imaging data processing and defect recognition technology based on deep learning

Authors: Dawen Yao; Peiwen Meng; Jinggang Xu; Shuqi Li

Addresses: School of Intelligent Equipment and Information Engineering, Changzhou Vocational Institute of Engineering, Changzhou, Jiangsu, China ' Huawei Technologies Co Ltd., Shenzhen, Guangdong, China ' School of Intelligent Equipment and Information Engineering, Changzhou Vocational Institute of Engineering, Changzhou, Jiangsu, China ' School of Intelligent Equipment and Information Engineering, Changzhou Vocational Institute of Engineering, Changzhou, Jiangsu, China

Abstract: This paper innovatively applies deep learning technology and uses a deep Convolutional Neural Network (CNN) to automatically extract key features from ultrasonic imaging data and perform defect recognition. The ultrasonic imaging data is denoised, normalised and data augmented, and a deep CNN model is constructed. Image features are automatically extracted through multi-layer convolution and pooling layers. The model is trained and optimised using the backpropagation algorithm and cross-entropy loss function. The trained model is used to realise real-time defect detection and precise positioning of new ultrasonic images. The defect classification and positioning model comparison experiment is compared with different CNN architectures such as Residual Network (ResNet), CBAM-CNN (Convolutional Block Attention Module CNN) and Hybrid CNN. The accuracy of the proposed method reaches 93.10%, and the detection speed is 832 images per second, which is significantly better than the detection precision and efficiency of other models.

Keywords: deep learning; industrial phased array ultrasonic imaging; defect recognition; deep convolutional neural network; data denoising; augmentation.

DOI: 10.1504/IJCAT.2026.151386

International Journal of Computer Applications in Technology, 2026 Vol.78 No.1, pp.39 - 52

Received: 14 Apr 2025
Accepted: 26 Jun 2025
Published online: 26 Jan 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article