微塑料和天然有机物混合物的自动分类使用深度学习模型
Seunghyeon Lee1, Heewon Jeong1, Seok Min Hong1
1Department of Civil Urban Earth and Environmental Engineering, Ulsan National Institute of Science and Technology (UNIST), UNIST-gil 50, Ulsan, 44919, Republic of Korea.
Water research
|October 19, 2023
概括
深度学习 (DL) 模型现在可以准确地分类水中环境中的微塑料 (MP),即使存在自然有机物 (NOM). 这种先进的方法绕过了耗时的预处理,提高了MP识别的准确性和客观性.
科学领域:
- 环境科学 环境科学
- 分析化学 分析化学
- 数据科学数据科学数据科学
背景情况:
- 在水生系统中微塑料 (MP) 的分类通常需要广泛的样品预处理,例如氧化以去除天然有机物 (NOM).
- 这些预处理步骤是耗时的,昂贵的,并且可以在光谱分析过程中引入由于主观操作员判断的错误.
- 准确和高效的MP识别对于了解它们对环境的影响至关重要.
研究的目的:
- 开发和评估深度学习 (DL) 模型,以提高与天然有机物 (NOM) 混合的微塑料 (MP) 的分类准确性.
- 评估一个带有空间注意力机制的卷积神经网络 (CNN) 的适用性,用于从拉曼光谱中分类MP.
- 将DL模型的性能与传统的拉曼光谱图书馆软件进行比较.
主要方法:
- 使用具有空间注意力机制的卷积神经网络 (CNN) 来分析微塑料与天然有机物 (MP-NOM) 混合物的拉曼光谱.
- 将DL模型的分类结果与传统拉曼光谱图书馆软件的分类结果进行了比较.
- 用梯度加权类激活映射 (Grad-CAM) 来研究DL模型训练和解释的关键光谱波段.
主要成果:
- 开发的DL模型实现了高分类准确率99.54%,显著超过传统的拉曼光谱图书馆软件 (31.44%).
- 格拉德-CAM分析证实了DL模型通过专注于突出的拉曼光谱峰值来识别MP的能力.
- 该模型在分类MP中表现出有效性,即使具有不那么突出的光谱特征,也表现出对峰值强度变化的稳定性.
结论:
- 深度学习为水生环境中的微塑料分类提供了一种强大,自动化和客观的方法.
- 拟议的DL模型消除了对繁的NOM预处理的需求,节省了时间和资源.
- 这项研究为推进微塑料分析和监测技术提供了一个有希望的方向.
相关概念视频
Polymer Classification: Architecture
Polymers are classified as linear or branched on the basis of their chain architecture. The polymer chains in linear polymers have a long chain-like structure with minimal to no branching at all. Even if a polymer features large substituent groups on the monomer, which appear as branches to the skeleton, it is not considered a branched polymer. A branched polymer contains secondary polymer chains that arise from the main polymer chain. The branching occurs when the polymer growth shifts from...
Polymer Classification: Stereospecificity
Polymerization generates chiral centers along the entire backbone of a polymer chain. Accordingly, the stereochemistry of the substituent group has a significant effect on polymer properties. Polymers formed from monosubstituted alkene monomers feature chiral carbons at every alternate position in the polymer backbone. Relative to the predominant orientation of substituents at the adjacent chiral carbons, the polymer can exist in three different configurations: isotactic, syndiotactic, and...
Classification and Mechanical Properties of Synthetic Polymers
Synthetic polymers are classified as elastomers, fibers, or plastics based on their crystallinity. Crystallinity, the degree of long-range order in the solid state, influences the mechanical properties (stretching or contracting) of elastomers. Elastomers are flexible polymers that can expand or contract easily upon the application of an external force. They have numerous crosslinks that pull them back into their original shape when stress is removed. Silicones, for instance, are highly elastic...
Bioplastics
Bioplastics derived from microbial processes present a sustainable alternative to conventional petroleum-based plastics. Among these, polyhydroxyalkanoates (PHAs), particularly polyhydroxybutyrates (PHBs), have emerged as prominent candidates due to their biodegradability and biocompatibility. These polymers are synthesized by a variety of bacteria, such as Cupriavidus necator and Pseudomonas putida, which naturally accumulate PHAs as intracellular carbon and energy reserves, especially under...
Microbial Bioremediation of Plastics
Polyethylene terephthalate (PET) is a synthetic polymer widely utilized in the packaging industry, particularly for bottles and containers. Due to its chemical stability and durability, PET accumulates in the environment, contributing significantly to plastic pollution. It comprises repeating units of terephthalic acid and ethylene glycol, resulting in a semi-crystalline structure that is resistant to natural degradation processes.A notable breakthrough in plastic biodegradation came with the...


