Unmanned aerial vehicle power grid inspection method and system based on SOC and MESH networking

Abstract

The invention relates to an unmanned aerial vehicle power grid inspection method and system based on SOC and MESH networking, and the method comprises the steps: carrying out the first training of a target detection model at a server side, and obtaining a first target detection model; loading the first target detection model to an inspection unmanned aerial vehicle; performing acceleration processing on the first target detection model based on a hardware and software acceleration module to obtain an accelerated second target detection model; based on the second target detection model, detecting an image obtained in the flight of the unmanned aerial vehicle to obtain detection result data; and training the second target detection model by using the detection result data to obtain a third target detection model. Through the MESH networking technology, a low-delay and high-resolution image transmission scheme is realized, and subsequent high-definition camera transmission is supported. The performance of the airborne SOC can realize the detection speed of 2-25 frames per second on the premise of meeting the precision, and the iterative model enables the detection to be more accurate.

Description

technical field [0001] This patent relates to the field of image detection. Specifically, it relates to a UAV power grid inspection method and system based on SOC and MESH networking technologies. Background technique [0002] Due to the particularity of high-altitude, long-distance and fixed line operations in power inspection, the use of drone inspections instead of manual inspections has been widely used in front-line teams. In the traditional drone inspection, the operator carries the drone to the operation point, and operates the take-off and landing of the drone through the aircraft remote control device. During the flight process, the inspection and data collection of key targets or routes are performed. [0003] Traditional inspection operations require the operator's attention to be highly concentrated. On the one hand, the control stability and safety of the UAV are ensured; abnormal situation. In recent years, artificial intelligence has developed rapidly, espe...

AI Technical Summary

Problems solved by technology

In addition, due to the extensive mileage of power grid lines and the limited signal coverage of network operators, 4G or wifi-based transmission methods are adopted. In complex landforms such as weak signals and mountainous areas, the limitation of signal coverage can easily cause the transmission failure of the system.

The computing equipment used in this patent has limitations, the algorithm supports less ecology, and relevant models are not accelerated, and the scalability needs to be improved.

In addition, due to the large number of tasks performed and the collection of a large amount of image data, the existing technology has not yet fully utilized automatic labeling to quickly read and iterate the model and update the data set

[0005] After analysis, the technical problems existing in the existing technical solutions are: 1) complex terrain and extensive transmission lines, the signal of network operators is limited or difficult to cover; Model Acceleration Optimization Tool

Images