DCC-CenterNet: A rapid detection method for steel surface defects

This paper presents an improved Convolutional Neural Network (CNN) architecture to recognize surface defects of the Calcium Silicate Board (CSB) using visual image information based on a deep learning approach. The proposed CNN architecture is inspired by the existing SurfNet architecture and is named SurfNetv2, which comprises a feature extraction module and a surface defect recognition module. The output of the system is the recognized defect category on the surface of the CSB. In the collection of the training dataset, we manually captured the defect images presented on the surface of the CSB samples. Then, we divided these defect images into four categories, which are crash, dirty, uneven, and normal. In the training stage, the proposed SurfNetv2 is trained through an end-to-end supervised learning method, so that the CNN model learns how to recognize surface defects of the CSB only through the RGB image information. Experimental results show that the proposed SurfNetv2 outperforms five state-of-the-art methods and achieves a high recognition accuracy of 99.90% and 99.75% in our private CSB dataset and the public Northeastern University (NEU) dataset, respectively. Moreover, the proposed SurfNetv2 model achieves a real-time computing speed of about 199.38 fps when processing images with a resolution of 128 × 128 pixels. Therefore, the proposed CNN model has great potential for real-time automatic surface defect recognition applications.

https://www.mdpi.com/1424-8220/20/16/4356/pdf

SurfNetv2: An Improved Real-Time SurfNet and Its Applications to Defect Recognition of Calcium Silicate Boards.

RDD-YOLO: A modified YOLO for detection of steel surface defects

An automatic workpiece surface defect detection system based on machine vision is proposed. In order to solve the difficulty of detecting surface defects of irregular workpieces, the system adopts seven shooting stations to collect pictures of workpieces from different angles. Then the pictures of each station are processed at the same time. Advanced image processing algorithms are proposed to detect defects on highly reflective metal surfaces. Finally, the processing results of all stations are combined together to determine whether the workpieces are qualified or not. Tested in real industrial environments, our proposed system achieves a detection accuracy of 99.7%. On average one hundred workpieces can be inspected per minute. 

Novel Automatic Workpiece Surface Defects Detection System

During the production and post-processing of complex metal parts, surface defects such as scratches, stains, and pits will inevitably occur, which will reduce the yield of metal parts and cause serious economic losses. Complex metal surfaces have complex surface texture characteristics, which leads to the difficulty of detection during the detection process. Therefore, this paper presents a defect inspection algorithm of metal surface based on machine vision. The proposed surface defect inspection algorithm first uses improved bi-dimensional empirical mode decomposition (BEMD)-based extracting algorithm to perform initial extracting of surface defects through filtering out complex textures on the metal surface, while retaining as much effective information as defects as possible. Then, the inspection algorithm applies Canny edge detection operator to detect the edge information of these defects. Finally, the final acquisition of defect edges can be achieved by connecting edge breakpoints and image filling operations. Experimental results on inspection of a variety of the surface defects on parts with metal surface are reported to show the performance of the defect inspection algorithm.

Meng Li

Papers

Defect Inspection Algorithm of Metal Surface Based on Machine Vision