環保專案成果報告查詢系統

專案計畫編號	EPA-110-A008
經費年度	110
計畫經費	2,900 千元
預算科目	固定污染源管制
專案開始日期	2021/01/09
專案結束日期	2021/12/10
專案主持人	江智偉
主辦單位	南區環境督察大隊
承辦人	王雅馨
執行單位	崑山科技大學
專案分類	大氣空氣
中文關鍵字	空氣污染、人工智慧物聯網、光達
英文關鍵字	Air pollutants, AIoT, LiDAR
協同主持人	許蕙琳
共同主持人
計畫聯絡人	江智偉
計畫聯絡電話	06-2051535#
計畫聯絡信箱	cwchiang@mail.ksu.edu.tw

中文摘要	本工作團隊利用科學儀器、物聯網系統以及AI智能空氣污染辨識系統等，以自動化、數位化、智能化和參數化，針對管道污染排放、污染空間分布、傳送與沉降、大氣條件和對附近空氣品質影響程度之關連性等，進行研究分析與驗證，已掌握部分高雄市大發工業區污染排放情形並進行有效污染防制與後續管制工作。相關工作執行與成果說明如下：工作執行方面，已完成建置9處AI智能污染辨識系統，進行工業區光達污染熱區掃描及管道連續監控各6場次，並進行8個月污染資訊收集、數據卷積神經網路（CNN）運算學習及影像辨識和污染資訊即時推播等，總時數達17,550小時。在污染事件分析與稽查管制成果方面，完成鎖定特定污染源管道，進行污染空間擴散、污染溯源、影響範圍和進行工業區空氣污染辨識和污染好發時間等分析，並彙整分析10件以上污染事件，協助污染稽查管制，糾舉違反空氣污染防制法之業者10件/次，有效杜絕污染情事。本計畫之AI智能空污辨識系統與相關稽查管制措施，已突破傳統之污染監測模式並有效提升稽查管制，且長期收集與彙整之空氣污染海量數據將可作為相關單位科學研究、環境評估和政策規劃之參考依據。
中文關鍵字	空氣污染、人工智慧物聯網、光達

中文摘要

本工作團隊利用科學儀器、物聯網系統以及AI智能空氣污染辨識系統等，以自動化、數位化、智能化和參數化，針對管道污染排放、污染空間分布、傳送與沉降、大氣條件和對附近空氣品質影響程度之關連性等，進行研究分析與驗證，已掌握部分高雄市大發工業區污染排放情形並進行有效污染防制與後續管制工作。相關工作執行與成果說明如下：工作執行方面，已完成建置9處AI智能污染辨識系統，進行工業區光達污染熱區掃描及管道連續監控各6場次，並進行8個月污染資訊收集、數據卷積神經網路（CNN）運算學習及影像辨識和污染資訊即時推播等，總時數達17,550小時。在污染事件分析與稽查管制成果方面，完成鎖定特定污染源管道，進行污染空間擴散、污染溯源、影響範圍和進行工業區空氣污染辨識和污染好發時間等分析，並彙整分析10件以上污染事件，協助污染稽查管制，糾舉違反空氣污染防制法之業者10件/次，有效杜絕污染情事。本計畫之AI智能空污辨識系統與相關稽查管制措施，已突破傳統之污染監測模式並有效提升稽查管制，且長期收集與彙整之空氣污染海量數據將可作為相關單位科學研究、環境評估和政策規劃之參考依據。

中文關鍵字

空氣污染、人工智慧物聯網、光達

類型	檔名	檔案大小	說明	最新版公開日期
期末報告	應用AI智能調查並查核大發工業區空品不良計畫(公開版).pdf	0MB	應用AI智能調查並查核大發工業區空品不良計畫	2022/1/1 上午 12:00:00

摘要	The objects of this project are to understand the pollution from exhausted stack, the spatial distribution, transmission and sedimentation of pollution, and the relationship with nearby air quality by using the scientific instruments, Internet of Things (IoT) and artificial intelligence (AI) identification system in Dafa Industrial Park. The implementation and results are described as follows: The AI intelligent air pollution identification system have been installed in 9 stations and executed 6 measurements of pollution by using 3-D LiDAR and Mini-CEMS(Continuous Emission Monitoring Systems), individually. The information of pollution has been collected and analyzed around 17,550 hours for convolutional neural network (CNN) computational learning, AI image identification and instant pollution notification. There are more than 10 pollution events found and assisted 10 inspection of pollution events which are against with Air Pollution Control Act etc.. This project has broken the traditional methodology with automation, intelligence, networking, and systemization to improve the effectiveness of environmental inspections, reduce the concentration of pollution, maintain the air quality, and protect people’s health. Furthermore, the massive data collected from this project will also be used as the reference for scientific research, environmental assessment and policy planning.
英文關鍵字	Air pollutants, AIoT, LiDAR

摘要

The objects of this project are to understand the pollution from exhausted stack, the spatial distribution, transmission and sedimentation of pollution, and the relationship with nearby air quality by using the scientific instruments, Internet of Things (IoT) and artificial intelligence (AI) identification system in Dafa Industrial Park. The implementation and results are described as follows: The AI intelligent air pollution identification system have been installed in 9 stations and executed 6 measurements of pollution by using 3-D LiDAR and Mini-CEMS(Continuous Emission Monitoring Systems), individually. The information of pollution has been collected and analyzed around 17,550 hours for convolutional neural network (CNN) computational learning, AI image identification and instant pollution notification. There are more than 10 pollution events found and assisted 10 inspection of pollution events which are against with Air Pollution Control Act etc.. This project has broken the traditional methodology with automation, intelligence, networking, and systemization to improve the effectiveness of environmental inspections, reduce the concentration of pollution, maintain the air quality, and protect people’s health. Furthermore, the massive data collected from this project will also be used as the reference for scientific research, environmental assessment and policy planning.

英文關鍵字

Air pollutants, AIoT, LiDAR