The most critical step in a line-structured light three-dimensional scan modeling system is to extract the centerline of the light stripe, but the interference of various environmental factors makes this extraction difficult. Several problems exists in a line-structured light streak image issues such as light spot interference, uneven distribution of light intensity, large differences in the width of the light bars, and complex background. This paper proposed a solution to overcome these problems. First, the structured light image is binarized using the Otsu method. Then, the improved density-based spatial clustering of applications with nose (DBSCAN) algorithm is used to retain the core points and remove the boundary and noise points. Finally, the core points are used as inputs to construct the graph data structure, and the shortest path search algorithm that fits the line-structured light streak image is used to obtain the center-line of the light streak. The experimental results show that the algorithm of this paper runs within 150 ms and the error is within 0.2 pixels. Moreover, this algorithm is applicable to various complex environments, meeting the requirements of real-time calculations, accuracy, and stability.

The probability of power transformer failure is extremely low, which leads to a great impact on further in-depth analysis results due to unbalanced data when processing transformer fault data. To solve these problems, this study processes and judges the unbalanced data using an confrontation neural network combined with an artificial neural network, uses the distributed acoustic wave sensing technology based on ultraweak fiber Bragg gratings to collect and analyze the data of the simulation site of the transformer built in a laboratory, and achieves good results on the collected transformer fault simulation data. This method has an important referential significance for developing the small sample fault identification system of the on-load transformer using confrontation generation network.

This paper proposes an artificial neural network-based fault detection and prediction model for on-load transformers using distributed fiber optic sensing technology. By artificially simulating the fault and normal operating states of transformers and using the k-means synthetic minority oversampling technique data expansion method, a small number of fault datasets can be limitedly expanded so that the amount of fault data can be consistent with that of normal data. Therefore, the expanded fault data and normal operation data can be input into the convolutional neural networks long short term memory identification model. Finally, the fault recognition rate can be increased to 100%, which has significant implications for the development of fault recognition systems for on-load transformers based on distributed fiber optic sensing technology.

Optical fiber microseismic monitoring technology is used to monitor and alert the microvibration events generated during production activities through observation and analysis with passivity and high reliability. Herein, the sensors are vertically installed on the side bolts along the roadway, and the monitoring substation is installed in the chamber. The sensors and the monitoring substation constitute a monitoring network through the laid optical cables. Besides, the simplex method is used to locate the seismic source. This method is free from divergence problems in the location calculation and is highly stable. Moreover, in this method, the solution of the partial derivative and inverse matrix is not required, which reduces the calculation amount and improves the calculation efficiency. Additionally, each sensor can use different wave velocities during the calculation based on the actual situation. The optical fiber microseismic monitoring system was installed in Shanxi Wuyang Coal Mine for preliminary monitoring and application, and the monitoring results were analyzed. The results show that the system can monitor mine activities and warn early, thereby playing a positive role in safe production.