Wireless Signal Propagation Prediction Based on Computer Vision Sensing Technology for Forestry Security Monitoring.

Area of Science:

• Geosciences
• Computer Science
• Electrical Engineering

Background:

• Forestry security monitoring requires accurate wireless signal propagation models.
• Topographic maps contain crucial data for predicting radio wave behavior.
• Existing methods for processing geographic information can be inefficient.

Purpose of the Study:

• To introduce Computer Vision (CV) sensing technology based on Convolutional Neural Networks (CNNs) for processing topographic maps.
• To predict wireless signal propagation models, specifically diffraction loss and shadow fading correlation distance.
• To enhance the accuracy and efficiency of radio propagation characteristic prediction in forestry security.

Main Methods:

• Utilized CV sensing technology with CNNs to process topographic map data.
• Generated two datasets for training CNNs on diffraction loss and correlation distance prediction.
• Implemented parallel processing in CNNs to predict multiple diffraction loss values simultaneously.

Main Results:

• Achieved high prediction accuracy for diffraction loss (95% error < 8.238%) and correlation distance (95% error < 6.423%).
• Demonstrated significant efficiency gains, predicting diffraction losses at 100 positions in approximately 6.28 ms.
• Showcased an average processing time per location point as low as 62.8 us.

Conclusions:

• The proposed CV sensing technology effectively processes geographic information for improved prediction accuracy and efficiency.
• CNNs enable a close coupling of prediction models with geographic data, enhancing wireless signal propagation modeling.
• This approach offers a superior method for forestry security monitoring through advanced geospatial data processing.