CottonNet-MHA: A Model for Detecting Cotton Disease Based on Deep Learning

Mostaque Md. Morshedur Hassan¹Asmita Ray²Munsifa Firdaus Khan Barbhuyan³Mudassir Khan^4,5*Bayan Alabdullah⁶Md. Faruqul Islam⁷Barga Mohammed Mujahid⁴

• ¹Brainware University, Kolkata, India
• ²Swami Vivekananda Institute of Science and Technology, Sonarpur, India
• ³School of Computing Science and Engineering and Artificial Intelligence, Kothri Kalan, India
• ⁴King Khalid University, Abha, Saudi Arabia
• ⁵Chitkara University, Rajpura, India
• ⁶Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia
• ⁷Global Institute of Management and Technology, Krishnanagar, India

India is an agro-based country. The major goal of agriculture is to produce disease-free healthy crops. For Indian agronomists, cotton is a profitable commercial and fiber crop, it is the world's second-biggest export crop after China. Cotton production is also affected in a negative way by high use of water, authority of soil erosion and the practice of using dangerous fertilizers and pesticides. The two greatest threats to the rapid growth of the crop are the sucking bugs and cotton diseases. The primary objective of this research is to build a model by implementing deep learning-based approaches to spot infections in cotton crops. Deep learning is used because of its exceptional results in classification and image processing tasks. In this study, five transfer learning architectures with pretrained weights, which are VGG16, VGG19, Inception V3, Xception, and MobileNet, are used to train our proposed model. We introduce CottonNet-MHA, an advanced deep learning-based architecture to detect pathological symptoms in cotton plants. The performance analysis is carried out on the developed model based on the conventional models and the results indicate that CottonNet-MHA dominates the conventional models with respect to its accuracy as well as efficiency in the detection of diseases. In our study, we have implemented Grad-CAM to enhance the trustworthiness and interpretability of the model. Additionally, we developed a web-based application using our proposed CottonNet-MHA architecture, which can effectively detect cotton diseases in real-world scenarios. The proposed model demonstrates high diagnostic accuracy for cotton diseases and opens new possibilities for the automatic detection of diseases in other plants as well.