Improved U-Net Drainage Pipe Defects Detection Algorithm Based on CBMA Attention Mechanism

Underground drainage pipelines are vital to urban infrastructure and play a crucial role in ensuring city safety. Existing methods for inspecting these pipelines struggle to achieve precise detection of pipeline defects in complex and low-visibility environments. Addressing the challenges of complex terrains and frequent defects in underground drainage pipelines, this paper integrates neural networks and deep learning to propose an intelligent detection algorithm based on an improved U-net with a CBAM attention mechanism. The algorithm uses the U-net model as a foundation for detecting pipeline defects, determining their presence and specific types; it employs VGG16 as the backbone for feature extraction to enhance model efficiency; and incorporates the CBAM attention module to increase detection accuracy. Experimental results show that the U-net model with the CBAM attention module performs well in detection tasks, achieving a high accuracy rate of up to 95.5%.