We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Automatic Identification and Classification of Marine Microplastic Pollution Based on Deep Learning and Spectral Imaging Technology
Summary
Researchers developed an AI system combining deep learning with multispectral imaging to automatically identify and classify marine microplastics, using a feature-selection method called ReliefF to reduce noise in complex ocean samples. The approach achieved high accuracy and offers a scalable solution for large-scale ocean microplastic monitoring that outperforms traditional manual inspection.
Marine microplastics pose a significant ecological and health risk due to their widespread sources and distribution.As a result, the rapid and accurate identification and classification of microplastics have become critical for marine environmental protection.Currently, traditional visual and microscope detection methods are inefficient and subjective.Some image-based recognition methods suffer from insufficient feature extraction capabilities, resulting in limited accuracy, while spectral-based techniques fail to effectively address data redundancy and noise, leading to poor classification performance in complex environments.To address these challenges, this study focuses on the development of an automatic recognition and classification technology for marine microplastic pollution using deep learning combined with spectral images.The research includes: proposing a feature extraction method for marine microplastics from multispectral images based on the ReliefF algorithm, which effectively selects features and removes redundant information; and developing a Conv-ReliefF-based recognition method for marine microplastics, integrating the feature learning ability of Convolutional Neural Networks (CNNs) with the feature selection advantages of the ReliefF algorithm.The innovation of this study lies in precisely extracting key features from multispectral images using the ReliefF algorithm to solve the problems of redundancy and noise interference in traditional feature extraction.By combining CNNs with the ReliefF algorithm, the Conv-ReliefF method balances feature learning depth and selective screening, thereby improving the accuracy and efficiency of microplastic recognition in complex marine environments.This approach provides technical support for large-scale marine microplastic monitoring.