環保專案成果報告查詢系統

專案計畫編號
經費年度	112
計畫經費	14,825 千元
預算科目	科技發展-淨零排放科技
專案開始日期	2023/03/01
專案結束日期	2023/12/15
專案主持人	朱仁佑
主辦單位	循環署循環處理組
承辦人	吳郁煌
執行單位	工業技術研究院
專案分類	科技計畫、資源循環
中文關鍵字	循環經濟、智能回收、塑膠再生
英文關鍵字	Circular Economy, Smart Sorting, Plastic Regeneration
協同主持人	江叡涵
共同主持人
計畫聯絡人	陳麗娟
計畫聯絡電話	03-5915164#
計畫聯絡信箱	anae_chen@itri.org.tw

中文摘要	循環經濟驅動下，各國陸續擬定各項限塑政策，國際各品牌大廠亦同步響應，承諾於2030年逐步降低並停止塑膠新料之使用，轉而以再生料取代，環境部以「扭轉廢棄資源走向循環未來」為政策主軸，在2023年2月預告「網際網路購物包裝限制使用對象及實施方式」草案，提升塑膠包材摻配25％以上再生料。目前全球除瓶罐及部分電器塑膠外，其他廢棄物的回收及再利用的比例低。為了提升塑膠包材回用率，本計畫將針對電子塑膠包材廢棄物循環鏈進行規劃，協助研擬「廢棄物智慧化自動分選及循環回收資源化平台規劃」，並以型態影像辨識技術、高通量光譜材質分選技術、電子塑膠包材再生技術作為三大核心技術，進行技術布局。本計畫起紇日為112年3月1日至12月15日，執行進度為100%。本計畫團隊已完成盤查國內外影像辨識、材質分選和電子塑膠包材再生循環技術研析。其中達成： 1.舉辦2場專家諮詢委員會議，如「廢塑料智慧化自動分選暨智能回收技術平台計畫-回收循環技術」及回收循環專家討論會議，瞭解國內環保業者對於電子塑膠包材回收分類需求及進展，研擬回收體系與再生處理物質通路，提出電子塑膠包材循環技術藍圖現況。 2.已完成智慧類型辨識技術，設計及布建影像辨識光學模組，並建置具有10227張影像的資料庫，可得到清晰回收物影像，進行特徵判讀。透過深度學習演算法模型類型辨識演算法開發，包含不同型態、紋路等差異性樣品，可辨識種類有5類(盛盤、泡棉、氣泡布、盒類、袋類)及不同顏色的包材，最終驗證結果高於目標90%，可達到16件/秒速度，精確度為96.4%。另盤點電子塑膠包材之材質，合計為315筆；辦理電子塑膠包材廢棄物之適性材質分類技術開發及建置資料庫，包含6100筆紅外線反射光譜數據。並透過特徵值演算法開發SAM快速光譜辨識技術，可分類PE、PET、PP、PS、纖維素、PVC六種包材材質，可達到18件/秒速度，平均精確度97.5%以上。 3.針對電子包材創新再應用技術開發部分，已完成單一塑膠材質平板包材(PP、PS、PET)之脫揮潔淨化及發泡緩衝材(EPP、EPO、EPS)減容之前處理及多層膜分離純化技術評估。另透過熔融造粒製程與基礎物性測試，輔以結構增強技術，擇定包材最佳處理方式，研擬創新分類及再生應用流程規劃，協助廢棄電子包材進入再生循環鏈前之再利用評估，以及包材分類再利用整合技術可行性評估。
中文關鍵字	循環經濟、智能回收、塑膠再生

中文摘要

循環經濟驅動下，各國陸續擬定各項限塑政策，國際各品牌大廠亦同步響應，承諾於2030年逐步降低並停止塑膠新料之使用，轉而以再生料取代，環境部以「扭轉廢棄資源走向循環未來」為政策主軸，在2023年2月預告「網際網路購物包裝限制使用對象及實施方式」草案，提升塑膠包材摻配25％以上再生料。目前全球除瓶罐及部分電器塑膠外，其他廢棄物的回收及再利用的比例低。為了提升塑膠包材回用率，本計畫將針對電子塑膠包材廢棄物循環鏈進行規劃，協助研擬「廢棄物智慧化自動分選及循環回收資源化平台規劃」，並以型態影像辨識技術、高通量光譜材質分選技術、電子塑膠包材再生技術作為三大核心技術，進行技術布局。本計畫起紇日為112年3月1日至12月15日，執行進度為100%。本計畫團隊已完成盤查國內外影像辨識、材質分選和電子塑膠包材再生循環技術研析。其中達成： 1.舉辦2場專家諮詢委員會議，如「廢塑料智慧化自動分選暨智能回收技術平台計畫-回收循環技術」及回收循環專家討論會議，瞭解國內環保業者對於電子塑膠包材回收分類需求及進展，研擬回收體系與再生處理物質通路，提出電子塑膠包材循環技術藍圖現況。 2.已完成智慧類型辨識技術，設計及布建影像辨識光學模組，並建置具有10227張影像的資料庫，可得到清晰回收物影像，進行特徵判讀。透過深度學習演算法模型類型辨識演算法開發，包含不同型態、紋路等差異性樣品，可辨識種類有5類(盛盤、泡棉、氣泡布、盒類、袋類)及不同顏色的包材，最終驗證結果高於目標90%，可達到16件/秒速度，精確度為96.4%。另盤點電子塑膠包材之材質，合計為315筆；辦理電子塑膠包材廢棄物之適性材質分類技術開發及建置資料庫，包含6100筆紅外線反射光譜數據。並透過特徵值演算法開發SAM快速光譜辨識技術，可分類PE、PET、PP、PS、纖維素、PVC六種包材材質，可達到18件/秒速度，平均精確度97.5%以上。 3.針對電子包材創新再應用技術開發部分，已完成單一塑膠材質平板包材(PP、PS、PET)之脫揮潔淨化及發泡緩衝材(EPP、EPO、EPS)減容之前處理及多層膜分離純化技術評估。另透過熔融造粒製程與基礎物性測試，輔以結構增強技術，擇定包材最佳處理方式，研擬創新分類及再生應用流程規劃，協助廢棄電子包材進入再生循環鏈前之再利用評估，以及包材分類再利用整合技術可行性評估。

中文關鍵字

循環經濟、智能回收、塑膠再生

類型	檔名	檔案大小	說明	最新版公開日期
期末報告	112AB008廢塑料智慧化自動分選暨智能回收技平台計畫成果報告(公開版).pdf	14MB		2024/2/21 上午 12:00:00

摘要	Under the circular economic initiative, countries are proposing plastic-limitation policies, and international brands are pledging to gradually replace new plastics with recycled materials by 2030. The Environmental Protection Administration Executive Yuan, R.O.C proposed a policy that focuses on "Reversing the Course of Waste toward a Circular Future." In February 2023, a proposal was drafted for "Internet Shopping Packaging Restrictions and Implementation," aimed at increasing the use of over 25% recycled materials in plastic packaging. Global recycling and reuse rates for waste, excluding bottles and some electronic plastics, remain low. To increase the reuse rates of plastic packaging, this project aims to plan a recycling chain for electronics plastic packaging waste and help develop a "Smart Automated Sorting and Circular Recycling Resource Platform." The project will use three core technologies: smart image recognition, smart material sorting, and plastic packaging regeneration. This project was started on March 1, 2023 and is to be concluded on December 15, 2023, with a completion rate of 100%. The project team successfully conducted a comprehensive review and analysis of domestic and international techniques for image recognition, material sorting, and the regeneration cycle of electronic plastic packaging. The key achievements of this project are as follows: (1) Two expert advisory committee meetings have been conducted, namely the "Smart Automated Sorting and Intelligent Recycling Technology Platform for Waste Plastics-Recycling Cycle Technology" and a discussion meeting with recycling experts. These meetings provided insights into the needs of the domestic waste-recycling industry and progress achieved in the classification and recycling of electronic plastic packaging. Subsequently, a recycling system was devised that can map out routes for the processing of regenerated materials and present the current state of the technological blueprint for the recycling of electronic plastic packaging. (2) We have completed the development of an intelligent recognition technology, designed and established an optical module for image recognition, and constructed a database with 10,227 images. As such, we could obtain clear images of recyclables that featured different packaging types. By developing a type recognition algorithm using deep learning models, materials with different forms, patterns, and other variations were discerned. Such packaging can be recognizably categorized into five types, namely trays, foam, bubble wrap, boxes, and bags, and with different colors. The final verification results exceeded the target of 90%, with an accuracy rate of 96.4% at the speed of 16 items/s. In addition, a survey of the materials used in electronic plastic packaging was conducted, totaling 315 entries. We also managed the development of suitable material-classification technology for electronic plastic packaging waste and established a database containing 6,100 infrared reflection spectroscopy data entries. By using the spectral angle mapper (SAM) in the rapid spectral-recognition technique with spectral characteristics, we could classify the packaging materials into PE, PET, PP, PS, cellulose, and PVC , achieving a speed of 18 items/s and maintaining an average accuracy of over 97.5%. (3) To develop innovative regeneration technologies for electronic packaging materials, we evaluated the devolatilization of flat packaging of a single plastic material (PP, PS, PET), performed pretreatment of foamed materials (EPP, EPO, EPS) for volume reduction, and used separation and purification technology to assess multilayer films. Furthermore, by using a melt-granulation process and testing fundamental physical properties, along with the use of structural-reinforcement techniques, we determined the optimal treatment methods for packaging. This involves the drafting of innovative, sorting, and regeneration-application-process planning techniques to assist in evaluating the regeneration before the electronic packaging material waste enters the recycling chain. Moreover, we assessed the feasibility of the integrated techniques for classifying and reusing packaging materials.
英文關鍵字	Circular Economy, Smart Sorting, Plastic Regeneration

摘要

Under the circular economic initiative, countries are proposing plastic-limitation policies, and international brands are pledging to gradually replace new plastics with recycled materials by 2030. The Environmental Protection Administration Executive Yuan, R.O.C proposed a policy that focuses on "Reversing the Course of Waste toward a Circular Future." In February 2023, a proposal was drafted for "Internet Shopping Packaging Restrictions and Implementation," aimed at increasing the use of over 25% recycled materials in plastic packaging. Global recycling and reuse rates for waste, excluding bottles and some electronic plastics, remain low. To increase the reuse rates of plastic packaging, this project aims to plan a recycling chain for electronics plastic packaging waste and help develop a "Smart Automated Sorting and Circular Recycling Resource Platform." The project will use three core technologies: smart image recognition, smart material sorting, and plastic packaging regeneration. This project was started on March 1, 2023 and is to be concluded on December 15, 2023, with a completion rate of 100%. The project team successfully conducted a comprehensive review and analysis of domestic and international techniques for image recognition, material sorting, and the regeneration cycle of electronic plastic packaging. The key achievements of this project are as follows: (1) Two expert advisory committee meetings have been conducted, namely the "Smart Automated Sorting and Intelligent Recycling Technology Platform for Waste Plastics-Recycling Cycle Technology" and a discussion meeting with recycling experts. These meetings provided insights into the needs of the domestic waste-recycling industry and progress achieved in the classification and recycling of electronic plastic packaging. Subsequently, a recycling system was devised that can map out routes for the processing of regenerated materials and present the current state of the technological blueprint for the recycling of electronic plastic packaging. (2) We have completed the development of an intelligent recognition technology, designed and established an optical module for image recognition, and constructed a database with 10,227 images. As such, we could obtain clear images of recyclables that featured different packaging types. By developing a type recognition algorithm using deep learning models, materials with different forms, patterns, and other variations were discerned. Such packaging can be recognizably categorized into five types, namely trays, foam, bubble wrap, boxes, and bags, and with different colors. The final verification results exceeded the target of 90%, with an accuracy rate of 96.4% at the speed of 16 items/s. In addition, a survey of the materials used in electronic plastic packaging was conducted, totaling 315 entries. We also managed the development of suitable material-classification technology for electronic plastic packaging waste and established a database containing 6,100 infrared reflection spectroscopy data entries. By using the spectral angle mapper (SAM) in the rapid spectral-recognition technique with spectral characteristics, we could classify the packaging materials into PE, PET, PP, PS, cellulose, and PVC , achieving a speed of 18 items/s and maintaining an average accuracy of over 97.5%. (3) To develop innovative regeneration technologies for electronic packaging materials, we evaluated the devolatilization of flat packaging of a single plastic material (PP, PS, PET), performed pretreatment of foamed materials (EPP, EPO, EPS) for volume reduction, and used separation and purification technology to assess multilayer films. Furthermore, by using a melt-granulation process and testing fundamental physical properties, along with the use of structural-reinforcement techniques, we determined the optimal treatment methods for packaging. This involves the drafting of innovative, sorting, and regeneration-application-process planning techniques to assist in evaluating the regeneration before the electronic packaging material waste enters the recycling chain. Moreover, we assessed the feasibility of the integrated techniques for classifying and reusing packaging materials.

英文關鍵字

Circular Economy, Smart Sorting, Plastic Regeneration