Research on Intelligent Monitoring Method of Bolt Looseness Based on Image Processing

Bolt connection In the process of autonomous driving of drones, the working performance of different types of bolts is of great significance. When the bolts of drones are interfered by the external environment, they will produce complex movements, which will seriously affect the safety of drones. In order to reduce the incidence of UAV accidents and ensure the safety of UAV flight, it is necessary to find a reasonable method [1-4] to conduct real-time status monitoring on the loosening of UAV bolts.

After high-altitude drones are disturbed by vibration, impact and high-temperature creep, the bolt structure of the drone is prone to slip, separate or fall off, which will affect flight safety. The traditional method [5-7] is to install sensors in the screw-fixed area Mainly, the cost is extremely high, and the sensor has poor anti-interference performance and poor working effect.

A machine vision image processing algorithms and kalman filtering algorithm: MVIP-Kalman (Machine Vision Image Processing Algorithms and kalman filtering algorithm: MVIP-Kalman) for bolt loosening monitoring for high-altitude drones is proposed. . The experimental results show that the machine vision image processing algorithm can accurately identify and track the bolt torus, the relative position error predicted by the filter has a fast convergence speed, and has a high monitoring accuracy, which meets the needs of high-altitude drone bolt looseness monitoring.

Visual feature selection after unmanned aerial vehicle bolt

When the bolt ring of the drone is detached, some areas of it will have more obvious identifiable features. The selected bolt ring detachment feature must meet the following specifications:

1) Simplification: Fewer bolt structure features are used to indicate the relative position of the bolt cap and rod to be monitored

2) Stability: The collection of bolt structure features must have robustness and feature uniqueness and non-ambiguity.