Sugarcane Yield Classification and Prediction Using Light weight Deep Network

Abstract

Sugarcane is the most important renewable commercial crops in India. The sugarcane farming and sugar industry are essential to the socio-economic development of rural communities by generating greater income and employment opportunities. The ability of decision-makers and planners to choose import or export strategies is based on the early detection and control of issues related to sugarcane yield indicators. In this manuscript, Sugarcane Yield Classification and Prediction Using Light weight Deep Network (SY-CP-LWDN) proposed. At first, the data are gathered via field and mill. Afterward, the data are fed to pre-processing using Triangle Filter (TF). Here the noise present in the data’s are reduced. Then these data’s are undergoes segmentation for segmenting canopy, leaf size and color. The segmented data’s are given for Adaptive Fuzzy Segmentation Algorithm (DAFSA) for segmenting canopy, leaf size and color of Sugarcane.  Then the segmented data’s are fed to the Feature extraction using Swin Transformer (ST) for extracting the features such as   Standardizing and Imputation. Finally classification is done using light weight deep network (LWDN). The classification results are grade1, grade2 and grade 3 of sugarcane yield. Simulation of the model done using python and performance metrics also examined. The performance metrics like accuracy, ROC used to analyze performance of proposed technique. The Performance of the proposed SY-CP-LWDN approach attains 24.11%, 27.12% and 32.73% high accuracy compared with existing methods such as Sugarcane yield prediction with data mining and crop simulation models (SYP-DM-CSM), Integration of RNN with GARCH refined by whale optimization algorithm for yield forecasting: a hybrid machine learning approach (IRNN-GARCH-WOA-HMA)  and An improved multilayer perceptron approach for detecting sugarcane yield production in IoT based smart agriculture(AIMP-DSYR-IoTSA), respectively.