LR-DETR: A Transformer-Enhanced Framework for Real-Time River Object Detection

The Detection Transformer (DETR) and the YOLO series have been at the forefront of advancements in object detection. The RT-DETR, a member of the DETR family, has notably addressed the speed limitations of its predecessors through a high-performance hybrid encoder that optimizes query selection. Building upon this foundation, we introduce the LR-DETR, a lightweight evolution of RT-DETR for river floating objects detection. This model incorporates the Advanced Screening Feature Path Aggregation Network (HS-PAN) to refine feature fusion via a novel Down-Top fusion path, significantly enhancing its expressive power. Further innovation is evident in the introduction of the Remainder Convolutional Network (RPCN) as the backbone, which selectively applies convolutions to key channels, leveraging the concept of residuals to reduce computational redundancy and enhance accuracy. The enhancement of the RepBlock with Conv3XCBlock, along with the integration of a parameter-free attention mechanism within the convolutional layers, underscores our commitment to efficiency, ensuring that the model prioritizes valuable information while suppressing redundancy. A comparative analysis with existing detection models not only validates the effectiveness of our approach but also highlights its superiority and adaptability. Our experimental findings are compelling: the LR-DETR achieves a 5% increase in mean Average Precision (mAP) at 0.5, a 25.8% reduction in parameter count, and a 22.8% decrease in GFLOPs compared to the RT-DETR algorithm. These improvements are particularly pronounced in the re-al-time detection of river floating objects, showcasing LR-DETR's potential in specific environmental monitoring scenarios.