IoT-Integrated Deep Learning Approach for Automated Disease Monitoring in Custard Apple Cultivation

The custard apple (Annona squamosa), also known as sugar apple, is an important medicinal plant widely cultivated in Bangladesh, India, Portugal, Thailand, Cuba, and the West Indies. However, it is highly susceptible to diseases such as anthracnose, black canker, leaf spot, diplodia rot, mealy bug, and foliar leaf spot. Early detection of these diseases is essential for sustaining crop health and yield. This paper presents a deep learning-based approach for disease prediction in custard apple plants using a hybrid convolutional neural network (CNN) architecture. Three pretrained models–MobileNetV2, ResNet152V2, and ConvNeXtTiny–enhanced with a squeeze-and-excitation mechanism for channel-wise attention, were combined through a fully connected network, achieving a classification accuracy of 99.15%. To extend this framework toward practical applications, we propose an IoT-enabled smart agriculture system that integrates image-based disease detection with environmental sensor data for continuous monitoring and real-time farmer alerts. The integration of CNN-based classification and IoT deployment highlights strong potential for sustainable agriculture and precision crop management, ensuring long-term productivity and resilience in custard apple cultivation.