研究目的
To develop a powerful analysis tool for rigorous pipeline inspection through the implementation of specific algorithms for precise delimitation of defective zones and reliable interpretation of defects, despite challenging conditions.
研究成果
The paper concludes that the Sauvola and Feng thresholding methods are effective for corrosion detection in pipeline inspection, with Feng method showing slightly better performance. However, expert validation is necessary to account for objects like welding joints that may influence interpretation results.
研究不足
The study acknowledges the presence of false positive indications and the need for expert validation to guide the choice of thresholding parameters. The complex nature of pipeline images requires supervised evaluation for realistic corrosion detection.
1：Experimental Design and Method Selection:
The study involves designing a motorized engine with an embedded camera controlled by FPGA technology for pipeline inspection and implementing image/video processing software for defect detection.
2：Sample Selection and Data Sources:
Videos and images acquired inside water transportation pipelines prone to corrosion are analyzed.
3：List of Experimental Equipment and Materials:
Includes a motorized engine with a digital camera, stepper motors, drivers, EPL camera, lighting sources, and batteries.
4：Experimental Procedures and Operational Workflow:
The motorized vehicle is controlled remotely to acquire video sequences inside the pipeline, which are then processed using Sauvola and Feng thresholding methods for corrosion detection.
5：Data Analysis Methods:
The performance of thresholding methods is evaluated based on the proportion of damaged areas and uniformity measures.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
