Scholarlink

The agriculture industry is hugely affected by the diseases found in Paddy and Wheat. To help in solving the impact of plant diseases, various technologies have been developed. However, Pre-Post-Harvest (PPH) diseases could not be predicted by them. Thus, aMulti-Task-level transfer Learning-based Softplus Deep Convolution Neural Network (MTL-SDCNN) model for predicting the PPH diseases of wheat and paddy is proposed. Primarily, the input images are collected and pre-processed. By employing the Gaussian Centre-Random Image Cropping and Patching (GC-RICAP), image patching is done in the pre-processed image. Next, feature extraction is carried out; in addition, by employing the Baytes Updated-Coati Optimization Algorithm (BU-COA), the important features are selected. Based on these features, the images are segmented as healthy and unhealthy using the Silhouette Score Distance modified Center Updated K-Means (SSD-CU-KM) algorithm. In the end, by employing the MTL-SDCNN, PPH diseases are predicted. As per the experimental outcomes, the proposed technique is more effective when analogized to the prevailing methodologies. The proposed method has 98.08% accuracy.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

MTL-SDCNN-Based Pre- and Post-Harvest Diseases Prediction of Wheat and Paddy Crops

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Sakshi Pandey,
  • Kuldeep Yogi,
  • Ayush Ranjan

摘要

The agriculture industry is hugely affected by the diseases found in Paddy and Wheat. To help in solving the impact of plant diseases, various technologies have been developed. However, Pre-Post-Harvest (PPH) diseases could not be predicted by them. Thus, aMulti-Task-level transfer Learning-based Softplus Deep Convolution Neural Network (MTL-SDCNN) model for predicting the PPH diseases of wheat and paddy is proposed. Primarily, the input images are collected and pre-processed. By employing the Gaussian Centre-Random Image Cropping and Patching (GC-RICAP), image patching is done in the pre-processed image. Next, feature extraction is carried out; in addition, by employing the Baytes Updated-Coati Optimization Algorithm (BU-COA), the important features are selected. Based on these features, the images are segmented as healthy and unhealthy using the Silhouette Score Distance modified Center Updated K-Means (SSD-CU-KM) algorithm. In the end, by employing the MTL-SDCNN, PPH diseases are predicted. As per the experimental outcomes, the proposed technique is more effective when analogized to the prevailing methodologies. The proposed method has 98.08% accuracy.