PipeDet: Pipe Detection in the Complex Structured Environment of Aeroengine Outer Profile

Pipeline detection is crucial for the assembly quality and operational safety of aeroengines, as precise gap measurement directly impacts engine performance, vibration characteristics, and long-term reliability under extreme flight conditions. However, in complex structural environments—characterized by dense component layouts, severe occlusions, and intricate spatial interferences—traditional detection methods often suffer from poor reliability, frequent missed detections, and false positives. To address these challenges, we propose PipeDet, an efficient end-to-end detection framework specifically designed for aeroengine pipelines. By seamlessly integrating the Dysample dynamic upsampling module and the novel (FPSM), PipeDet reliably extracts fine-grained 2D pipeline features even in highly cluttered environments. The detected 2D bounding boxes and segmentation masks are then projected into 3D space via a calibrated multi-view system, enabling accurate pipeline instance segmentation. Following this, high-precision gap measurement is achieved through rapid feature reconstruction. Experiments on a dedicated hybrid aeroengine pipeline dataset demonstrate that PipeDet significantly outperforms state-of-the-art detectors in both accuracy and speed. This advancement effectively addresses the challenges of accurate and efficient aeroengine detection and gap calculation in complex environments.