Accurate Estimation of LeafWetness Duration using Deep Learning Based Time-Resolution Technique

Abstract

Plant diseases are one of the leading causes of loss of crops. Monitoring the plant growth conditions could help prevent the spread of diseases and ensure healthy agricultural productivity. A leaf wetness sensor (LWS) is used to detect wetness on the surface of the leaves. Leaf wetness duration (LWD) is an essential parameter that governs the possibility of the occurrence of diseases in plants. The sensor system used in this work for monitoring the environmental conditions transmits data every 30 minutes due to power consumption considerations. To improve the time-resolution of the signal and accurate estimation of LWD, a deep learningbased architecture is proposed. The data from LWS has been used to train the architecture and the results obtained have been compared with the existing methods to analyze its performance. Using the proposed architecture, the signal from LWS was bench-marked with the commercially available Phythos-31. The timeresolution of data has been improved from � 30 minutes to � 4 minutes. The average SNR values for three test signals with improved time-resolution have increased from 19.36dB to 22.16dB. Similarly, the average RMSE of the estimated LWD values after time-resolution improved from 0.55 to 0.20. During the experimental analysis, it has been observed that the proposed architecture estimates the LWD values better compared to the other time resolution techniques.