Surface Defect Detection of Mining Automation Equipment Based on Convolutional Neural Networks

Abstract

Surface defect detection of mining automation equipment is crucial for creating a safe operating environment. This paper analyzes the commonly used target detection techniques and selects a target detection algorithm based on Faster-RCNN to construct a model for surface defect detection in mining automation equipment. In the acquisition of surface defects, the Laplace operator and homomorphic filtering are used to enhance and sharpen the image. The texture defect dataset and metal surface defect dataset are also selected to make the VOC2007 dataset, and the Faster-RCNN network model is utilized for training to obtain the surface defect detection model of mining automation equipment. The model defect detection comparison results show that the Faster-RCNN model has an mAP value of 76.1%, which is the model with the highest detection accuracy. The result confirms the effectiveness of the method presented in this paper and enhances the accuracy of the model’s detection.

Keywords

• target detection
• faster-rcnn
• image acquisition
• mine automation equipment
• 97p10