Design and Implementation of an Inspection Robot for Crack Detection in Flooded Pipes

The pipeline used to supply drinking water to Andean communities is a crucial component of the water management project in mining complexes that aims to preserve local resources. However, the degraded state of the pipeline poses significant environmental, social and economic risks. To overcome these challenges, the present work proposes the development of a teleoperated robot for the inspection of flooded pipelines. The vehicle features an IP68 protection system to ensure underwater operation, along with measurement and analysis capabilities to assess the pipeline profile. In addition, it incorporates real-time video transmission and visual inspection aids through a graphical interface. Effective communication is ensured within a distance of 600 meters between the vehicle and the command post with a latency of less than 100 ms. Furthermore, an important aspect of this work is the successful segmentation of cracks and fissures using a hybrid CNN-RNN architecture, trained on a dataset comprising online repository images and images captured inside the pipeline.