Overview / Usage

PROBLEM STATEMENT :

Continuous cutting down of trees called deforestation may have a huge impact on our ecosystem and is one of the major sources for climate change, desertification, soil erosion, fewer crops, flooding, increased greenhouse gases in the atmosphere, and a host of problems for Indigenous people. Deforestation may lead to sudden/drastic climatic change and also worsen the environment . By this project we can predict the forest cover increase/decrease to be addresed . And this also finds the barren land in the specific areas and predict the soil and temperature so that specific tree which can be survived/grow well in that soil/temperature can be planted there. So our aim is to curb deforestation and to increase number of trees using modern age tech like AI and oneAPI tools. Through this project there will be 10-15% increase in number of trees per year. If this project is used efficiently it may also lead to huge increase in percentage of trees.

​

Methodology / Approach

This project includes three pre trained models trained in Intel oneAPI one to detect the forest areas , barren lands etc.. which is trained using CNN algorithm .The latitude and longitude of the barren land will be sent to anther model called soil-temp-predictor which will predict the soil and temperature of that area. the predicted output(soil and temp) is sent to another model called tree-predictor which will predict the tree which can survive well at that soil and temperature .So that we can send drones with seed balls to that area for seeding. The seed ball is composed of necessary nutrients required by the seed to grow initially. The drones deliver these seed balls to the area which has been detected using the model .The models are then integrated to android studio to covert it into android app .It uses Google Earth API where users can select a specific area , the image of that area is sent to the model so that it detects forest area, barren land ,soil ,temperature and suitable trees to be grown.

Technologies Used

Technologies used here is Intel oneAPI toolkit in Intel DevCloud.

LIbraries used are:

oneDAL library

oneDNN library

daal4py library

Intel-Tensorflow kernel ,oneAPI kernel ,python(oneAPI) kernel

and ML libraries like tensorflow , keras etc..

​

Documents and Presentations

Repository

https://github.com/vidhyavasan07/Curb-Deforestation-using-oneAPI.git