2025/01/01 by Qiangbao Ouyang, Yu Fang, Xintian Liu +4
Engineering · Mathematics · Physics and Astronomy · #Advanced Measurement and Detection Methods #Advanced Optical Sensing Technologies #Artificial intelligence #Computer science #Computer vision #Distance measurement #Fusion #Image (mathematics) #Image fusion #Laser #Laser ranging #Line (geometry) #Materials science #Mathematics #Optical Systems and Laser Technology #Optics #Physics #Ranging #Telecommunications #Transmission (telecommunications) #Transmission line
paper · doi:10.1109/tim.2025.3580837
openalex publication_date 2025/01/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/06/26
Existing sag measurement systems are often hindered by complex workflows and heavy reliance on manual assistance. An intelligent sag measurement method, integrating image data and laser ranging technology, is proposed. Based on this method, an intelligent sag measurement system is developed to enable automatic coordinate collection and sag calculation. The method uses a laser rangefinder to measure distances on the transmission line, which are converted into three-dimensional coordinates using angular relationships from a spatial position model. A catenary model is then applied to fit the data and calculate the sag value. In the system, a sag measurement algorithm is designed to automatically determine horizontal and pitch rotation parameters. Horizontal rotation angles are calculated by uniformly controlling rotation distances based on the number of measurement points. For pitch rotation, an AutoML-optimized BP neural network is constructed, using laser distances and image pixel differences as inputs. Model performance is evaluated using an absolute error threshold. The experimental results show that the proposed pitch angle prediction model achieves a coverage rate of 99.040% within the error tolerance range. The average MAE of the sag intelligent measurement system is 0.111 m, the average RMSE is 0.140 m, and the average standard deviation is 0.129 m. The measurement time is improved by 24.390% compared to a total station.