Two-Fold Spoiled Onion Detection using Soft Computing and IoT

Abstract

With the advancement of technology and the dependency of people on phones, it is important to come up with solutions involving technology. Using traditional storage methods, farmers can inhibit the spoilage of onions. But, in some situations, people may fail to notice the spoiled onions and in such a scenario they can depend on technology involving some deep learning algorithms and sensors. In the existing techniques, the system which consists of IoT framework alone faced many challenges because sometimes it may predict the data wrongly due to environmental conditions and leading it to an inefficient technique to detect onion spoilage. To overcome this kind of challenge, it is a must that technology like image processing should be included. This paper discusses the model that was developed using Google Colab IDE, which is based on image processing. Combining the segmentation and object extraction process has improved the image features, as it discards the background and other unnecessary things around the main object in our application. CNN model has got 87% accuracy, this shows a good result after evaluation. After this image processing segment, the work continues with the IoT framework that senses the parameters of onions using esp8266 & sensors, and it displays the stages of spoilage on LCD. Through this system, farmers and retail sellers can get early information about the spoilage of onions by accessing the real-time values through the web page.