A novel multi-model cascade framework for pipeline defects detection based on machine vision

MEASUREMENT（2023）

引用 1|浏览2

暂无评分

摘要

•An image acquisition system is designed to capture pipeline surface images;•A SRGAN modified with a self-attention mechanism is used to enhance image quality;•An IVGG16 classifier is improved to effectively enhance the detection performance;•A Mask R-CNN model is developed to locate defects in panorama images.

查看译文

关键词

Pipeline defects detection,Multi-model cascade,Machine vision,Imbalanced dataset,Low-quality images

AI 理解论文

溯源树

样例

生成溯源树，研究论文发展脉络

Chat Paper

正在生成论文摘要