Published Paper Details:

Dry and wet segregation using CNN with robotic arm is used to segregate the waste into dry and wet class. The aim of this research work is to segregate the trash between dry and wet using image classification technologies and deep learning algorithm for detecting trash. This research paper will help to improve trash management systems. Convolutional Neural Networks (CNN) are based on the transfer learning architecture, were developed to search for trash objects in an image and separate dry and wet items from the trash objects, respectively. In this reseach paper, we are using dataset of trashNet where we train and test the dataset of trash to classify the class between dry and wet. Using TrashNet image dataset we achieved great performance. Then the system was trained and tested on real images taken by the user in the intended usage environment. Using the image data, the first CNN achieved a preliminary 84.97% accuracy to identify dry and wet items on an image dataset of assorted trash items. Finally, a robotic arm controlled by the microcontroller(Raspberry Pi) is used to pick up the garbage and places it in the bin. As this model segregates waste automatically without any human intervention, this model can be very useful in handling waste which can pose a huge risk on human life.