Machine Learning-Driven Optimization of Plant Growth Through Dynamic Climate Regulation

This paper presents a novel intelligent greenhouse monitoring and management system integrating advanced deep learning techniques, including Convolutional Neural Networks (CNN), Deep Neural Networks (DNN), and Generative Adversarial Networks (GAN). The proposed model enhances precision agriculture by automating dynamic climate control, early plant disease detection, and compensating for limited real-world datasets through synthetic data augmentation. Multi-modal sensor data—temperature, humidity, soil moisture, $\mathbf{p H}$, and $\mathbf{C O}_{\mathbf{2}}$ levels—are continuously collected and analysed by a DNN to optimize environmental parameters and improve energy efficiency. Concurrently, a CNN is employed for accurate classification of plant diseases based on leaf images. Because of dataset limits, GANs construct high-quality artificial images that help expand the training corpus and make the model more reliable. Experiments show that using the integrated system can detect diseases with up to 96% accuracy, reduces labour intervention by more than $70 \%$ and saves $18 \%$ on power use. Real-time sensing, AI making decisions and data enrichment enable the system to successfully support smart, sustainable greenhouse farming on a bigger scale.