Intelligent detection of facility layout compliance of barrier-free toilet based on improved YOLOv8

Aiming at the problems of low efficiency, strong subjectivity and difficulty in quantitative evaluation of traditional manual detection method in barrier-free toilet facilities layout compliance inspection, this study proposes an intelligent detection method based on improved YOLOv8. The FasterNet backbone network is introduced to reduce the computational complexity of the model, and the BiFormer attention mechanism is combined to enhance the ability of the model to extract key facility features, and the Focal-EIoU loss function is used to optimize the training process of the bounding box regression. The experimental results show that the improved algorithm achieves a precision of 91.3%, a recall of 89.7%, a FPS of 68, and a FLOPs of 4.1G. In the measurement of key indicators of layout compliance, the distance error is controlled between 3 mm and 8 mm. This method improves the automation level and accuracy of compliance detection of barrier-free toilet facilities layout, and provides effective technical support for standardized acceptance and digital management of building barrier-free environment.