A massive amount of plastic waste has presented an immense management challenge.This escalating ecological damage,coupled with the detrimental effects of plastics infiltrating the marine food web,poses a significant t...A massive amount of plastic waste has presented an immense management challenge.This escalating ecological damage,coupled with the detrimental effects of plastics infiltrating the marine food web,poses a significant threat to human livelihoods.To combat this,there is a call for the development of plastic detection algorithms using remote sensing data.Here we tested a new index,referred to index_(MP),to detect clusters of floating macro plastics in the ocean using satellite imagery.The index_(MP)was applied to convolution high-pass filtered(3×3)Sentinel 2 Level 1 C images,showing the potential to reduce atmospheric interference and enhance the object edges,thereby improving the clarity of detection.In the analysis,we used three scatter plots to identify and assess plastic pixels.To differentiate the common features of plastic from non-plastic objects,the Sentinel 2 bands 5,8,and 9 were plotted against index_(MP)calculated and convolution high-pass filtered Level 1 C(CHPIC)images.The plastic pixels,clustering in the three scatter plots,showed positive‘X’,i.e.,CHPIC image value and‘Y’,i.e.,each band 5,8,and 9 reflectance values,along with a CHPIC image value exceeding 0.05.Using the index_(MP)and scatter plot analysis,we identified plastic pixels containing 14%or more plastic bottles.Detection of other types of plastics,such as fishing nets and plastic bags,required pixel proportions greater than 50%.Hence,plastic bottles were notably responsive even at a low pixel fraction.We further explored the classification of plastic and non-plastic objects by analyzing reed(plant)pixels;the differentiation between plastic and reed was conducted in the band 5 and 9 scatter plots.展开更多
基金Supported by the Guangdong Special Support Program for Key Talents Team Program(No.2019BT02H594)the PI Project of Southern Marine Science and Engineering Guangdong Laboratory(Guangzhou)(No.GML2021GD0810)the Major Project of National Social Science Foundation of China(No.21ZDA097)。
文摘A massive amount of plastic waste has presented an immense management challenge.This escalating ecological damage,coupled with the detrimental effects of plastics infiltrating the marine food web,poses a significant threat to human livelihoods.To combat this,there is a call for the development of plastic detection algorithms using remote sensing data.Here we tested a new index,referred to index_(MP),to detect clusters of floating macro plastics in the ocean using satellite imagery.The index_(MP)was applied to convolution high-pass filtered(3×3)Sentinel 2 Level 1 C images,showing the potential to reduce atmospheric interference and enhance the object edges,thereby improving the clarity of detection.In the analysis,we used three scatter plots to identify and assess plastic pixels.To differentiate the common features of plastic from non-plastic objects,the Sentinel 2 bands 5,8,and 9 were plotted against index_(MP)calculated and convolution high-pass filtered Level 1 C(CHPIC)images.The plastic pixels,clustering in the three scatter plots,showed positive‘X’,i.e.,CHPIC image value and‘Y’,i.e.,each band 5,8,and 9 reflectance values,along with a CHPIC image value exceeding 0.05.Using the index_(MP)and scatter plot analysis,we identified plastic pixels containing 14%or more plastic bottles.Detection of other types of plastics,such as fishing nets and plastic bags,required pixel proportions greater than 50%.Hence,plastic bottles were notably responsive even at a low pixel fraction.We further explored the classification of plastic and non-plastic objects by analyzing reed(plant)pixels;the differentiation between plastic and reed was conducted in the band 5 and 9 scatter plots.
