This project is a mobile application with a web service to let the farmers detect the diseases of the plants in early stages by taking pictures of the leaves and send it to the server in order to classify the disease.
The targeted users are the Egyptian farmers.
This is why the application should be user friendly and easy to use as much as possible. The early versions will not work on all plants while the database is not populated with many types of plants. The result processing time will depend on the computational power of the server.
Pre-processing is the first phase in the system and it should extract the leaf from the image to decrease the time of processing in the classifier. And background of each image will be removed using background subtraction technique. Then feature extraction will be applied in the server, the purpose of feature extraction is to reduce the original dataset by measuring certain features or properties of each image such as texture, color and shape. In order to recognize and identify healthy and infected leaf, measure several numbers of features in acquired image, to be later use for classification. The final step of this system is the classification phase, neural networks algorithm will be applied for classifying the plant leaf image to any of the following states, healthy or infected. The inputs of this stage are training dataset; the outputs will be the decision that determine type of input image (healthy of infected). The final product is a mobile app that will take picture of the leaf and then will send it to the web server which will do all the mentioned processing.
Our project has qualified to the pre-final stages in Microsoft and itworx competitions and the final in EMC.