Computer Vision for Smart Farming and Sustainable Agriculture

Developments in satellite technology, remote sensors and drone technologies are mushrooming. These developments yield volumes of high quality scene images that require effective processing for intelligent farming applications. The recent deep learning technologies can leverage these opportunities to fuse computer vision and artificial intelligence in farming. This encompasses the big data phenomena and huge volumes of data that are captured, processed and applied for decision-making. This paper aims to give insights on the integration of computer vision for smart farming in-order to attain sustainable agriculture. Using a structured approach, this research proposes a computer vision technique for crop image feature characterization that applies in the determination of the crop's health status. To achieve this, a deep convolutional network is applied for image feature extraction and representation, and then these features are fed to the support vector-learning machine for training and subsequent image interpretation. From the experimental results, it is evident that the proposed technique generates superior visual interpretation results of scene images as compared to other methods in literature. It follows that the Global food security and agricultural sustainability can be attained through ICT enabled solutions that are integrates and works together a phenomenon referred to as smart farming.