AutomaticRobotic Action Towards Plastic Collection

Significant wet and dry wastes produced by plastic. Almost all industries use plastic extensively. In India, plastic waste is typically dumped in natural resources, which dramatically increases environmental pollution. Humans frequently segregate plastic by hand, which is a poor solution. The concepts of deep learning (DL) and the internet of things (IOT) are combined in the suggested approach. IOT is utilized to pick up discovered plastic objects while a deep learning approach is utilized to detect plastic objects. Real-time object detection is crucial for locating plastic trash. As a feature extractor, Mask R-CNN and Resnet 101 are employed. Goal detection model Mask R-CNN is combined with Resnet 101 residual network, which speeds up neural network training. Resnet 101 utilises the residual subsystem to help the model converge. Prediction confidence scores are used to set an appropriate benchmark to reduce false positives because plastic things have various levels of transparency. Robotic arms are used to collect plastic trash that has been detected. Three motors are used to power the robotic arm. Two DC motors move the mobile robot forward and backward, and a third DC motor with a gripper is used to pick up the plastic object. Arduino Uno R3 manages the robotic arm's motions. The outcomes of the tests demonstrate that the suggested approach satisfactorily tracked different kinds of plastic particles and picked them up with a robot. The success rate of detection ranges from 85% to 93%. Different kinds of risks may be found in the future and employed for commercial and societal purposes.