ABSTRACT
Due to the technology development in the field of drones, multispectral analysis and remote sensing promise its support in the agriculture-related activities in terms of analysis and testing variables. This method shows the possibility of efficient plantation using photogrammetry of multispectral images and analyzing the plant indices. From the past decade in agriculture research, imaging data has grown exponentially using drones and satellites. This has given an opportunity for farmers to use the technology in the field of farming. They can read plant indices obtained from multispectral analysis with higher efficiency to make decision within the field. Different methodologies are being developed across the world for the detection of various crop diseases and nitrogen requirement prediction. Following objectives need to be achieved: (1) developing model to collect geo-located images using drone integrated with multispectral camera and (2) obtaining plant indices using image processing techniques. High-quality imaging allows farmers to make decisions precisely and much earlier before the disease spreads in the field. The important goal of developing multispectral algorithms is to reduce the pesticide and nitrogen usage. In the proposed method, plant indices are obtained from image processing that is done using data obtained from multispectral camera integrated to UAVs. These plant indices which are geo-located maps help in taking decision regarding agriculture activities. This can be done without compromising the image quality. The proposed algorithm helps in obtaining plant indices like NDVI, NDRE through detection, segment and classification using image processing techniques. It may help in taking proper decision regarding agriculture activities like reducing the usage of pesticides and nitrogen requirement prediction and to improve outcomes. Multispectral analysis is helpful in taking decision regarding agriculture activity. The primary objective is to create an ecosystem including drone integrated with multispectral camera, GPS and acquisition of geo-located images. However, image resolution limits the altitude and area of land coverage. An efficient automated method is given to detect disease and recommend nitrogen using the proposed algorithm.
