Crop diseases can cause significant yield losses and economic damage, the classification and estimation of the crop diseases in the early stage of disease helps in reducing these losses. Crop disease classification is a critical component of precision agriculture, which aims to improve crop yields and quality by using data-driven crop management methods. In our study, we make use of the ConvNeXT model architecture for the classification of rice crop disease. Two different ConvNeXT models are created by training with thermal and RGB images dataset respectively. By combining the output of both model accuracy scores, classification is done. The estimation of the disease in the affected rice leaves is done using the computer vision technique which estimates the percentage of the leaf area affected by the crop disease. The proposed method for classification achieved an accuracy of 98.7% for thermal dataset and 93.6% for the RGB dataset.