PREDICTION of AMOUNT of PESTICIDES and DISEASE IDENTIFICATION in FRUITS USING IOT and DEEP LEARNING

Abstract

Akshaya A

The use of pesticides,steroids and fertilizers has tremendously increased which has been found to have serious effect on human body giving rise to deadly diseases such as cancer, genetic disorders,etc., The problem statement is to develop a device that would identify the disease and predict whether the fruit is affected or not by the pesticide. A total of 120 classes consisting of different varieties of fruits with the training set size of nearly 60000 images and test set size of nearly 20000 images with an image size of 100x100 pixels (per image) were taken into account and the training were given with the help of CNN and SVM algorithms. The information whether the pesticides are present or not is decided by Maximum Residue Limits(MRL). of test fruits were found to belong in  the  threshold range of which proved  that  those  fruits  were  affected with pesticides. The percentage of pesticides affected, the information about the disease is sent to the application of the consumer’s phone thereby a real time regular monitoring of pesticides is possible.