環境資源報告成果查詢系統

110年度屏東縣精進砂石疏濬專區污染科技監控計畫

中文摘要 本計畫自110年4月1日起即進行建立道路髒污AI辨識功能相關作業等,期末報告統計執行成果至110年11月25日止,各工作項目均依規定完成。鄰近屏北砂石疏濬專區揚塵空品監測站,110年7月至10月PM10平均監測濃度25.2μg/m3,較109年同期平均監測濃度39.0μg/m3,下降35.3%。環保局執行民眾滿意度調查,110年持正面評價80.5%,較109年持正面評價76%提升4.5%,顯示砂石場稽查管制計畫再納入本計畫智能監控與派遣,除達到污染源頻率降低,如本年度UAV空中稽查,查獲違規7件,較去年違規率降低50%;CCTV查獲違規123件次,較去年降低63%;砂石車核查696輛,符合比例96%,較去年提升1%,更獲民眾的肯定。透過與七河局聯合管制河川揚塵,包含高屏溪裸露地改善作為、加強疏濬工區、疏濬車輛之污染管制、協助河川揚塵預警應變及加強專用道洗掃等,降低河川揚塵發生。以下就本計畫執行摘要成果進行說明。 本計畫建立道路髒污AI辨識功能CCTV揚塵監控系統連線10處,系統辨識共計10,069件次,真陰性9,504件次,真陽性393件次,整體系統辨識準確度98%。另系統通報10處CCTV道路髒污事件計512件次,包含真陽性有393件次,經人工複查偽陽性有119件次,偽陽性率1.2%,透過人工複查,偽陰性有53件次,偽陰性率12%。應變端方面,藉由本年度本計畫執行屏北砂石疏濬專區,完成建立道路髒污AI辨識功能及AI判斷道路髒污啟動因子預警通報及應變改善,降低人力成本,縮短污染排放時間,共計發佈污染通報393次,完成改善364次,改善率93%,另自10月份開始單月改善率為100%。通報污染改善364次,改善砂石疏濬專區聯外道路揚塵污染長度為2184公里,依據環保署街道揚塵洗掃計算TSP削減量30.1公噸,PM10削減量5.7公噸。透過屏北砂石疏濬專區鄰近土庫國小揚塵空品監測站,110年7月至11月與109年同期測值比較,降低幅度達31%。 運用砂石場洗車台運行智能監控方面,整體通報89次,39次為無砂石車進出,50次為空車離開砂石場未使用洗車台,橫向聯繫空污子計畫加強輔導砂石場空車進洗車台,經空污子計畫執行現場輔導及蹲點查核後,2家砂石場自10月份開始自律後已無空車未進洗車台情形,全掌控洗車設備使用率,使用率由原本85%提升至100%,PM10削減量1.5公噸/年。 輔導里港鄉及高樹鄉道路聯合認養洗街車設置GPS行車紀錄器,納入智能監控系統及定時未移動通報派遣,計通報54次,6月至11月份GPS智能監控通報數有逐月下降趨勢,11月份2處通報洗街車逾時未移動次數僅各1次,主要因素為駕駛用餐及生理需求,並不影響整體全日聯外道路認養洗街完整度。整體推動輔導道路認養洗街車設置GPS行車紀錄器,納入智能監控系統,獲得管理道路認養洗街車正常化運作成效,提升2處認養專區年度執行里程數計2,156km,TSP削減量為30公噸,PM10削減量為5.6公噸。 比較驗證屏北砂石疏濬專區主要聯外道路,各區域街塵負荷改善情形,數據顯示計畫設置AI監控系統前4月-6月及設置後7月-11月街塵負荷平均值變化,4條砂石車主要聯外道路街塵負荷,7月-11月皆較4月-6月為低,街塵負荷降低率介於23~41 %。另3處河川揚塵監測站分別於5月至6月完成設置。里港朧祥河濱公園監測站,監測日起至10月監測值皆低於預警值(100 μg/m3),11月1日至11月24日有44次小時監測值大於預警值,經人員現場巡查無河川揚塵情形,通報砂石場稽查管制計畫,協助聯繫砂石場加強執行場內及聯外道路灑水,持續監測濃度並無提升之情形。新園鄉萬丹第一抽水廠監測站與新園國小監測站,監測日起至10月監測值皆低於預警值(100 μg/m3),11月1日至11月24日有15次小時監測值大於預警值(100 μg/m3),經人員現場巡查,無河川揚塵發生。此外,設置空氣品質微型感測器與鄰近高屏溪流域河川揚塵監測站平行比對結果顯示,6處揚塵測點之空氣品質微型感測器回推後的PM10濃度均貼近揚塵測站之PM10監測值,儘管仍具一定程度的差異,然在濃度變化上兩者趨勢一致。 空氣品質微型感測器砂石疏濬專區空品監控運用分析方面,整體PM10測值,在12小時均值除少數幾天較高外,其他監測值在50μg/m3以下,顯示砂石專區污染管制是有成效的。分析砂石場運作時段(上午7時至下午19時)PM10小時監測共計有5次濃度超過100μg/m3,調閱鄰近CCTV揚塵監控畫面,主要為受道路乾燥有輕微車行揚塵影響,經空污子計畫通報認養聯外道路砂石業者加強道路沖洗後,PM10測值在2個小時內即降低,顯示微型感測器仍可運用在砂石疏濬專區空品之預報與即時性啟動污染防制措施之預防。 透過本計畫科技監控,包含CCTV、AI辨識道路髒污及通報系統、砂石場洗車台運行監控、聯外道路聯合認養委外洗街灑水車GPS監控、新設河川揚塵監測站強化高屏溪流域揚塵監測網及微型感測器之運用,並與經濟部水利署第七河川局及橫向聯繫空污子計畫,提早進行預防性灑水作業與部署應變能量,分析河道揚塵時PM10監測最大小時濃度平均值為116 μg/m3,較109年事件日PM10最大小時濃度平均值(396 μg/m3)降低70%。
中文關鍵字 AI辨識、PM10

基本資訊

專案計畫編號 經費年度 110 計畫經費 9480 千元
專案開始日期 2021/04/01 專案結束日期 2021/12/31 專案主持人 周俊利
主辦單位 屏東縣政府環境保護局 承辦人 蕭智元 執行單位 康廷工程顧問企業有限公司

成果下載

類型 檔名 檔案大小 說明
期末報告 110年度屏東縣精進砂石疏濬專區污染科技監控計畫.pdf 48MB 110年度屏東縣精進砂石疏濬專區污染科技監控計畫

Advanced Technology Monitoring Plan for Pollution in Sand and Gravel Dredging Zone of Pingtung County in 2021

英文摘要 This project initiated from April 1, 2011 to carry out the work related to the establishment of the AI identification function of road pollution. The implementation results of the final report are collected until November 25, 2011, and all task items have been completed in accordance with the contract. The fugitive dust monitoring station adjacent to the Pingbei Sand and gravel dredging zone, the average concentration of PM10 from July to October of 2011 was 25.2 μg/m3, which was 35.3% lower than the 39.0 μg/m3 in the same period of 2010. The results of public satisfaction survey carried out by the Environmental Protection Bureau of Pingtung County Government (PTEPB) showed that 80.5% of the public evaluations were positive in 2021, an increase of 4.5% compared with 76% in 2010. It demonstrates that the inspection and control plan on sand and gravel field integrated with the intelligent monitoring and dispatch of this plan, in addition to reducing the occurrence frequency of pollution sources, is more affirmed by the public. The UAV aerial inspection, 7 violations were found in 2021, a 50% decrease compared with those in 2020. Violations of 123 cases were seized by CCTV, a 63% decrease from last year. The 696 sand and gravel vehicles were inspected, with a compliance rate of 96%, an increase of 1% from last year. The occurrence frequency of river fugitive dust is reduced by cooperation with the 7th Rivers Management Office of Water Resources Agency to control the river fugitive dust, including the improvement of bare riverbed along the Gaoping River, enhancing the pollution control of dredging vehicles around the dredging work area, assisting in the early warning and response of river fugitive dust, and strengthening the washing and sweeping of dedicated roads, etc. The execution results of this plan are summarized as the followings. This project established 10 connections of the CCTV fugitive dust monitoring system with the AI identification function of road pollution. The system identified a total of 10,069 cases, including 9,504 true negatives and 393 true positives. The overall system identification accuracy is 98%. In addition, the 10 CCTV system reported 512 cases of road pollution incidents. There were 393 true positives, of which 119 false positives after manual recheck, and a false positive rate of 1.2%. There were 53 false negatives through the manual recheck, the false negative rate was 12%. On the response measures executed in the Pingbei sand and gravel dredging zone this year, through the establishment of the AI identification function of road pollution and initiating factors of the road pollution judged by AI to issue the early warning notifications, as well as the improvement of the response, the overall labor costs were reduced and the emission duration of pollutants was also shortened. A total of 393 pollution notifications were made, of which 364 improvements were completed, and the improvement rate was 93%, and a monthly improvement rate increased to 100% since October in 2021. The improvement of pollution notifications was completed 364 times, and the length of road fugitive dust pollution around the sand and gravel dredging zone ameliorated was 2,184 kilometers. According to the guidelines of street dust removed by washing and sweeping operations released by Taiwan Environmental Protection Administration (EPA), the estimated reduction amounts of TSP and PM10 were 30.1 and 5.7 tons, respectively. Through the air monitoring station for fugitive dust in the Tuku Elementary School adjacent to the Pingbei sand and gravel dredging area, the measured values from July to November in 2011 was compared with the same period in 2010, and the reduction rate was 31%. In terms of the intelligent monitoring of the operation of the car wash in the sand and gravel yard, the overall notifications were 89 times, 39 times for no sandstone vehicles entering or leaving the gravel yard, and 50 times for the empty vehicles leaving the gravel field without using the car washing platform. The task team notified the sub-project of air quality control to strengthen the guidance of the empty vehicles which should enter the car washing platform before leaving the gravel yard. After the implementation of on-site coaching and on-site inspection of the sub-projects of air quality control, the two sand and gravel fields have no empty cars not entering the car washing platform since they started self-discipline in October, and the utilization rate of car washing equipment has been fully controlled and the utilization rate increased from 85% to 100%. The PM10 reduction amount was estimated about 1.5 tons/year. Guiding the street washing vehicles jointly provided by Ligang Township and Gaoshu Township to set up the GPS driving recorders, incorporating into the intelligent monitoring system and non-movement notifications within a specified period, with a total of 54 notifications. From June to November, the number of GPS intelligent monitoring reports showed a downward trend month by month. In November, only 1 time each was notified in two places for non-movement of the street washing vehicle over the specified period. The main reasons were eating a meals and physiological needs of the driver, which did not affect the overall full-day progress of street washing operation. The overall promotion of the guidance on road maintenance of street washing vehicles to install GPS driving recorders, incorporating into the intelligent monitoring system, has achieved the regular operations and management of the street washing vehicles, and the annual road length of 2 maintenance areas has been increased by 2,156 km, and the reduction amounts of TSP and PM10 were 30 and 5.6 tons, respectively. Compare and verify the improvement of street dust loading from April to June before the plan to install the AI monitoring system and from July to November after the installation on the main roads around the Pingbei sand and gravel dredging zone. The data showed the average of street dust loading from July to November for the 4 main roads used by sandstone vehicles were lower than those from April to June, and the reduction rate of street dust loading was between 23% and 41%. In addition, 3 river dust monitoring stations were set up between May and June. The Luxiang Riverside Park Monitoring Station in Ligang Township, from the day of operation of monitoring station to October, the monitoring values were all lower than the early warning value (100 μg/m3), but from November 1 to November 24, there were 44 hourly monitoring values greater than the early warning value. The task team of this plan inspected on-site to confirm that there was no river fugitive dust occurred, then notified the inspection and control team for sand and gravel field to assist the contact to the operator of sand and gravel field to strengthen the implementation of watering on the roads inside and outside the site, and continued to monitor the situation that the concentration did not increase. The Wandan No. 1 Pumping Plant Monitoring Station and Xinyuan Elementary School Monitoring Station in Xinyuan Township, from the day of operation of monitoring station to October, the monitoring values were all lower than the early warning value (100 μg/m3), but from November 1 to November 24, there were 15 hourly monitoring values greater than the early warning value. The task team of this plan inspected on-site to confirm that there was no river fugitive dust occurred. In addition, the parallel comparison results of the air quality micro-sensors and the nearby river fugitive dust monitoring stations along the Gaopingxi River Basin showed that the PM10 concentrations of the air quality micro-sensors at the 6 fugitive dust monitoring points were all close to those of the fugitive dust monitoring stations. Although there were still some variations in the monitoring values, the two trends were consistent in terms of the concentration variations. In terms of the application of air quality micro-sensors on the air quality monitoring in the sand and gravel dredging zone, the values of 12-hour average were higher only in a few days among the overall PM10 measurement values, the other monitoring values were below 50 μg/m3, indicating that the pollution control in the sand and gravel zone is effective. During the operation period of the sand and gravel field (from 7:00 to 19:00), a total of 5 times of hourly PM10 monitoring values were higher than 100 μg/m3. Checked the nearby CCTV fugitive dust monitoring record, the slight fugitive dust was mainly caused by the vehicles running on the dry road. Through the notifications of the sub-projects of air quality control to the operators, the PM10 measurement values decreased in 2 hours after enhancing the road washing of outreach roads around the sand and gravel field, indicating that the micro sensor can still be used in the forecast of air quality in the sand and gravel dredging zone and the immediate activation of pollution prevention and control measures. Through the technology monitoring of this project, including CCTV, AI identification of road pollution and notification system, operation monitoring of car washing platform in gravel yard, GPS monitoring of street washing vehicles, new river fugitive dust monitoring stations to strengthen the dust monitoring network and application of micro-sensors along the Gaopingxi watershed, and cooperation with the 7th Rivers Management Office of Water Resources Agency and close contact with the sub-projects of air quality control to carry out preventive watering operations and deploy response capacity in advance. The average of monitoring maximum hourly PM10 concentration during the occurrence of river fugitive dust is 116 μg/m3 in 2021, which is 70% lower than those (396 μg/m3) on the event days of 2020.
英文關鍵字 AI identification function, PM10