Deep Learning Implemented Visualizing City Cleanliness Level by Garbage Detection

Abstract

In an urban city, the daily challenges of managing cleanliness are the primary aspect of routine life, which requires a large number of resources, the manual process of labour, and budget. Street cleaning techniques include street sweepers going away to different metropolitan areas, manually verifying if the street required cleaning taking action. This research presents novel street garbage recognizing robotic navigation techniques by detecting the city’s street-level images and multi-level segmentation. For the large volume of the process, the deep learning-based methods can be better to achieve a high level of classification, object detection, and accuracy than other learning algorithms. The proposed Histogram of Oriented Gradients (HOG) is used to features extracted while using the deep learning technique to classify the ground-level segmentation process’s images. In this paper, we use mobile edge computing to process street images in advance and filter out pictures that meet our needs, which significantly affect recognition efficiency. To measure the urban streets’ cleanliness, our street cleanliness assessment approach provides a multi-level assessment model across different layers. Besides, with ground-level segmentation using a deep neural network, a novel navigation strategy is proposed for robotic classification. Single Shot MultiBox Detector (SSD) approaches the output space of bounding boxes into a set of default boxes over different feature ratios and scales per attribute map location from the dataset. The SSD can classify and detect the garbage’s accurately and autonomously by using deep learning for garbage recognition. Experimental results show that accurate street garbage detection and navigation can reach approximately the same cleaning effectiveness as traditional methods.