Urban Street Cleanliness Assessment using Machine Learning

In order to solve the increasing amount of Object detection is a core part in computer vision, with applications spreading throughout the process of constructing a smart city, city managers always invest much energy and capital in cleaning street trash because of the random emergence of street trash, As deep network solutions become increasingly deep and convoluted, they tend to be constrained by the quantity of training data. Keeping this in view, to drive developments in analysing and interpreting images, Open CV or Google AI has made the Open Images dataset public. Open Images is an extension of the tradition initiated by PASCAL VOC, Image Net and COCO, but now in an unprecedented scale. In this project we to execute the Hence, visual street cleanliness evaluation is especially significant. Yet, current assessment methods have some evident shortcomings, i.e., collecting street garbage data is not automated and street cleanliness data is not real-time best performing algorithm for automatically detecting objects. Lastly, the results are integrated into the street cleanliness calculation model to finally visualize street cleanliness levels, which brings convenience to city managers in organizing clean-up staff effectively.