一种融合移动边缘计算和深度学习的城市街道垃圾检测和清洁度评估方法
|
摘要
针对近年来城市化进程的加快,城市街道垃圾的随意出现给市政部门及时清扫带来巨大困难.在移动、云技术和物联网飞速发展的背景下,论文研究并提出了一种融合移动边缘计算和深度学习的城市街道垃圾检测和清洁度评估方法,通过安装在车辆上的高分辨率摄像机也称为"移动站"和手持的移动设备进行街景图象收集,利用边缘服务器临时存储并处理街景图象信息,然后通过城市网络把这些数据传输到云中心,利用深度学习技术识别街道垃圾类别以及对垃圾数量计数,并且将这些结果引入到基于层次的街道清洁度评估框架当中,最终可视化街道清洁度等级,为城市市政管理者有效安排清理人员提供方便.文章最后基于南京市江宁区的街道图像,可视化了江宁区街道清洁度等级.实际应用初步验证了其可行性和可用性.
In response to the acceleration of the urbanization process in recent years,the random appearance of urban street garbage has brought great difficulties to the municipal departments in timely cleaning. Under the background of the rapid development of mobile,cloud technology and the Internet of Things,this paper studies and proposes an approach for urban street garbage detection and cleanliness assessment fusing mobile edge computing and deep learning. The high resolution cameras installed on vehicles collect street viewimages,which is also called " mobile station". Using edge servers to store temporarily and extracting street viewimage location information,then transmitting these data to the cloud center for analysis through city networks. At the same time,deep learning technology is used to identify street garbage categories and counting the number of garbage. These results are incorporated into the street cleanliness calculation framework to ultimately visualize street cleanliness levels,which provides convenience for city municipal managers to arrange clean-up personnel effectively. At the end of the paper,the street cleanliness level of Jiangning District was visualized based on the street image of Jiangning District in Nanjing. The practical application verifies the feasibility and availability of the system preliminarily.
In response to the acceleration of the urbanization process in recent years,the random appearance of urban street garbage has brought great difficulties to the municipal departments in timely cleaning. Under the background of the rapid development of mobile,cloud technology and the Internet of Things,this paper studies and proposes an approach for urban street garbage detection and cleanliness assessment fusing mobile edge computing and deep learning. The high resolution cameras installed on vehicles collect street viewimages,which is also called " mobile station". Using edge servers to store temporarily and extracting street viewimage location information,then transmitting these data to the cloud center for analysis through city networks. At the same time,deep learning technology is used to identify street garbage categories and counting the number of garbage. These results are incorporated into the street cleanliness calculation framework to ultimately visualize street cleanliness levels,which provides convenience for city municipal managers to arrange clean-up personnel effectively. At the end of the paper,the street cleanliness level of Jiangning District was visualized based on the street image of Jiangning District in Nanjing. The practical application verifies the feasibility and availability of the system preliminarily.
引文
[1] Balchandani C,Hatwar R K,Makkar P,et al. A deep learning framework for smart street cleaning[C]. IEEE Third International Conference on Big Data Computing Service and Applications,2017:112-117.
[2] Begur H,Dhawade M,Gaur N,et al. An edge-based smart mobile service system for illegal dumping detection and monitoring in San Jose[C]. IEEE International Conference on Smart City and Innovation,2017.
[3] Shi Wei-song,Sun Hui,Cao Jie,et al. Edge computing-an emerging computing model for the Internet of everything era[J]. Computer Research and Development,2017,54(5):907-924.
[4] Wu Xi-bo,Yang Zai-gao. The concept of smart city and future city development[J]. Urban Studies,2010,17(11):56-60.
[5] Aguilera U,Pe1a O,Belmonte O,et al. Citizen-centric data services for smarter cities[J]. Future Generation Computer Systems,2017,76(11):234-247.
[6] Badii C,Bellini P,Cenni D,et al. Analysis and assessment of a knowledge based smart city architecture providing service APIs[J]. Future Generation Computer Systems,2017,75(10):14-29.
[7] Zygiaris S. Smart city reference model:assisting planners to conceptualize the building of smart city innovation ecosystems[J].Journal of the Knowledge Economy,2013,4(2):217-231.
[8] Hefnawy A,Bouras A,Cherifi C. Integration of smart city and lifecycle concepts for enhanced large-scale event management[C]. IFIP PLM'15 International Conference on Product Lifecycle M anagement,2015:687-697.
[9] Santur Y,Karak9se E,Karak9se M,et al. Deep learning based artificial manager for smart city[C]. 5th International Conference on Advanced Technology&Sciences,2017:197-201.
[10] Polishetty R,Roopaei M,Rad P. A next-generation secure cloudbased deep learning license plate recognition for smart cities[C].15th IEEE International Conference on M achine Learning and Applications(ICM LA),2016:286-293.
[11] Zhou Shu-ren,Tan Feng,Zeng Dao-jian. Human action recognition based on multi-size pooling of convolutional neural network[J].Journal of Chinese Computer Systems,2017,38(8):1893-1898.
[12] Mittal G,Yagnik K B,Garg M,et al. SpotGarbage:smartphone app to detect garbage using deep learning[C]. Proceedings of the 2016ACM International Joint Conference on Pervasive and Ubiquitous Computing,2016:940-945.
[13] Rad M S,Kaenel A V,Droux A,et al. A computer vision system to localize and classify wastes on the streets[C]. International Conference on Computer Vision Systems,2017:195-204.
[14] Borozdukhin A,Dolinina O,Pechenkin V. Approach to the garbage collection in the“smart clean city”project[C]. 4th IEEE International Colloquium on Information Science and Technology(CiSt),2016:918-922.
[15] Ren S,He K,Girshick R,et al. Faster R-CNN:towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and M achine Intelligence,2017,39(6):1137-1149.
[16] Girshick R. Fast R-CNN[C]. Proceedings of the IEEE International Conference on Computer Vision,2015:1440-1448.
[17] Zeiler M D,Fergus R. Visualizing and understanding convolutional networks[C]. European Conference on Computer Vision,Springer,Cham,2014:818-833.
[18] Krizhevsky A,Sutskever I,Hinton G E. Imagenet classification with deep convolutional neural networks[C]. Advances in Neural Information Processing Systems,2012:1097-1105.
[19] Jang Y C,Jain P,Tolaymat T,et al. Characterization of pollutants in Florida street sweepings for management and reuse[J]. Journal of Environmental M anagement,2009,91(2):320-327.
[20] Sevilla A,Rodríguez M L,García-Maraver,et al. An index to quantify street cleanliness:the case of Granada(Spain)[J]. Waste M anagement,2013,33(5):1037-1046.
[21] López I,Gutiérrez V,Collantes F,et al. Developing an indicators plan and software for evaluating street cleanliness and waste collection services[J]. Journal of Urban M anagement,2017,6(2):66-79.
[3]施巍松,孙辉,曹杰,等.边缘计算:万物互联时代新型计算模型[J].计算机研究与发展,2017,54(5):907-924.
[4]巫细波,杨再高.智慧城市理念与未来城市发展[J].城市发展研究,2010,17(11):56-60.
[11]周书仁,谭凤,曾道建.多尺寸池化卷积神经网络的人体行为识别研究[J].小型微型计算机系统,2017,38(8):1893-1898.
1http://cleanstreetsla. com/clean-streets-challenge/2017.
[2] Begur H,Dhawade M,Gaur N,et al. An edge-based smart mobile service system for illegal dumping detection and monitoring in San Jose[C]. IEEE International Conference on Smart City and Innovation,2017.
[3] Shi Wei-song,Sun Hui,Cao Jie,et al. Edge computing-an emerging computing model for the Internet of everything era[J]. Computer Research and Development,2017,54(5):907-924.
[4] Wu Xi-bo,Yang Zai-gao. The concept of smart city and future city development[J]. Urban Studies,2010,17(11):56-60.
[5] Aguilera U,Pe1a O,Belmonte O,et al. Citizen-centric data services for smarter cities[J]. Future Generation Computer Systems,2017,76(11):234-247.
[6] Badii C,Bellini P,Cenni D,et al. Analysis and assessment of a knowledge based smart city architecture providing service APIs[J]. Future Generation Computer Systems,2017,75(10):14-29.
[7] Zygiaris S. Smart city reference model:assisting planners to conceptualize the building of smart city innovation ecosystems[J].Journal of the Knowledge Economy,2013,4(2):217-231.
[8] Hefnawy A,Bouras A,Cherifi C. Integration of smart city and lifecycle concepts for enhanced large-scale event management[C]. IFIP PLM'15 International Conference on Product Lifecycle M anagement,2015:687-697.
[9] Santur Y,Karak9se E,Karak9se M,et al. Deep learning based artificial manager for smart city[C]. 5th International Conference on Advanced Technology&Sciences,2017:197-201.
[10] Polishetty R,Roopaei M,Rad P. A next-generation secure cloudbased deep learning license plate recognition for smart cities[C].15th IEEE International Conference on M achine Learning and Applications(ICM LA),2016:286-293.
[11] Zhou Shu-ren,Tan Feng,Zeng Dao-jian. Human action recognition based on multi-size pooling of convolutional neural network[J].Journal of Chinese Computer Systems,2017,38(8):1893-1898.
[12] Mittal G,Yagnik K B,Garg M,et al. SpotGarbage:smartphone app to detect garbage using deep learning[C]. Proceedings of the 2016ACM International Joint Conference on Pervasive and Ubiquitous Computing,2016:940-945.
[13] Rad M S,Kaenel A V,Droux A,et al. A computer vision system to localize and classify wastes on the streets[C]. International Conference on Computer Vision Systems,2017:195-204.
[14] Borozdukhin A,Dolinina O,Pechenkin V. Approach to the garbage collection in the“smart clean city”project[C]. 4th IEEE International Colloquium on Information Science and Technology(CiSt),2016:918-922.
[15] Ren S,He K,Girshick R,et al. Faster R-CNN:towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and M achine Intelligence,2017,39(6):1137-1149.
[16] Girshick R. Fast R-CNN[C]. Proceedings of the IEEE International Conference on Computer Vision,2015:1440-1448.
[17] Zeiler M D,Fergus R. Visualizing and understanding convolutional networks[C]. European Conference on Computer Vision,Springer,Cham,2014:818-833.
[18] Krizhevsky A,Sutskever I,Hinton G E. Imagenet classification with deep convolutional neural networks[C]. Advances in Neural Information Processing Systems,2012:1097-1105.
[19] Jang Y C,Jain P,Tolaymat T,et al. Characterization of pollutants in Florida street sweepings for management and reuse[J]. Journal of Environmental M anagement,2009,91(2):320-327.
[20] Sevilla A,Rodríguez M L,García-Maraver,et al. An index to quantify street cleanliness:the case of Granada(Spain)[J]. Waste M anagement,2013,33(5):1037-1046.
[21] López I,Gutiérrez V,Collantes F,et al. Developing an indicators plan and software for evaluating street cleanliness and waste collection services[J]. Journal of Urban M anagement,2017,6(2):66-79.
[3]施巍松,孙辉,曹杰,等.边缘计算:万物互联时代新型计算模型[J].计算机研究与发展,2017,54(5):907-924.
[4]巫细波,杨再高.智慧城市理念与未来城市发展[J].城市发展研究,2010,17(11):56-60.
[11]周书仁,谭凤,曾道建.多尺寸池化卷积神经网络的人体行为识别研究[J].小型微型计算机系统,2017,38(8):1893-1898.
1http://cleanstreetsla. com/clean-streets-challenge/2017.