Drone-based Detection and Geo-Mapping of Wastes, Weeds and Diseases in Plants using Deep Learning

It has been recognised in recent years how detrimental pollution has been to the environment, as evidenced by global warming and a decline in the standard of living in developing nations. When it comes to pollution, plastic pollution is one of the most serious issues that humanity has created. Apart from the fact that plastics are non-biodegradable, they also reduce plant yields and deplete soil fertility. To avoid this situation, waste management education is required. It is critical to eliminate these contaminants prior to their ability to seep deeply into the soil. Weeds are unwelcome plants that grow in close proximity to other plants. They not only supplant healthy plants in terms of development, but also feed on them, robbing them of nutrients and resulting in a significant decrease in agricultural production. Another factor complicating agricultural production is the presence of plant diseases. Scientists are engaged in a variety of endeavours, including forecasting illnesses, predicting their potential, and identifying diseases before they become fatal.Our new method employs a first-of-its-kind strategy that involves mapping a region using a drone’s vector and then detecting map errors using the UDP stream from the drone’s camera, a first in the industry. This data is analysed in real time using Nvidia Deep Stream. To train our data-set, we will use a custom-built Sequential Convolution Neural Network.