We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Improving cross-flight plastic litter segmentation with ConvNeXt V2 U-Net for UAV SWIR hyperspectral imagery
AI summary Read the abstract
Scientists have improved a drone-based imaging technique that spots plastic litter in the environment (like on beaches or in waterways) even when lighting and weather conditions change between flights—a common problem that used to trip up earlier detection systems. Better plastic-tracking tools like this matter because they help researchers and cleanup crews find and remove pollution before it breaks down into microplastics, which can enter water supplies, food chains, and ultimately our bodies. While this study focuses on environmental monitoring rather than direct health testing, catching plastic pollution earlier and more reliably is a key step toward reducing our overall exposure to
Plastic litter represents a persistent form of environmental pollution, and accurate spatial monitoring is essential for ecosystem assessment, pollution mitigation, and evidence-based management. UAV short-wave infrared hyperspectral imaging can detect plastic materials with high spatial and spectral detail, but its operational use remains limited by strong variability between acquisitions collected on different flight days, where changes in illumination, exposure, weather, and surface background introduce cube-specific radiometric differences that reduce cross-flight generalisation. This study addresses plastic litter segmentation as both a remote sensing and data harmonisation problem by standardising each hyperspectral cube using its own spectral statistics before segmentation, reducing acquisition-specific bias without relying on training-set statistics. On this harmonised input, we introduce a compact ConvNeXt V2-based U-Net adapted to the 60 retained SWIR bands, with pyramid pooling to incorporate multi-scale contextual information and CBAM decoder refinement to improve background rejection. The framework is evaluated on a public multi-flight UAV SWIR benchmark and compared with PSPNet, SegFormer, DeepLabV3, U-Net, UNet++, and our previous attention-gated residual U-Net. Compared with our previous work, the proposed model improves the mean Dice score from 64.6% to 78.2% and the mean IoU from 49.0% to 64.5%; under a matched retraining protocol that equalises the training-data budgets of the two models, the mean leave-one-cube-out IoU rises from 0.37 to 0.69. The improvement is especially important under difficult acquisition conditions, where the previous model was more affected by overexposure, reduced spectral contrast, and background confusion. The incremental architectural study further shows that pyramid pooling and CBAM refinement improve the balance between plastic detection and background rejection, increasing precision, specificity, AUC-ROC, AUC-PR, and IoU compared with the ConvNeXt V2 baseline on a challenging out-of-distribution cube. Multi-seed analysis on the development cubes shows low sensitivity to random initialisation, indicating that variation across these runs is driven more by acquisition conditions than by initialisation. Overall, the proposed framework provides a compact and transferable approach for UAV-based hyperspectral plastic litter monitoring under realistic cross-flight conditions.
More Papers Like This
Attention-Gated U-Net for Robust Cross-Domain Plastic Waste Segmentation Using a UAV-Based Hyperspectral SWIR Sensor
AI summary Read the abstract
A deep learning model using attention-gated U-Net architecture achieved up to 96.8% accuracy in segmenting plastic waste in UAV-captured shortwave infrared hyperspectral imagery across varied real-world environments. This approach significantly outperforms classical machine learning methods and enables scalable, automated monitoring of plastic litter across natural ecosystems using remote sensing.
Classification of marine plastic debris using hyperspectral imaging and band selection: A patch-based and pixel-based fusion approach.
AI summary Read the abstract
Researchers developed an uncertainty-aware fusion of a lightweight hyperspectral convolutional neural network with a random forest classifier to identify five common marine plastic polymers, achieving 97% classification accuracy on full hyperspectral images and maintaining 90% accuracy with only six selected spectral bands—reducing computational demand by more than 20-fold for potential real-time drone-based monitoring.
Advances and challenges in plastic detection: a critical review of studies employing hyperspectral data
AI summary Read the abstract
This review pulls together 29 studies on using satellites, drones, and special cameras that read light beyond what our eyes can see (hyperspectral imaging) to spot plastic pollution in the environment. The technology works better for some plastics (like common packaging materials) than others, and still struggles when plastic is wet, degraded, or mixed with other debris—meaning better tools are needed before we can reliably track where plastic waste ends up in our air, water, and soil. This matters because the more we can map plastic pollution at its source, the better equipped we are to reduce the microplastics that eventually make their way into our food, water, and bodies.
Aerial detection of beached marine plastic using a novel, hyperspectral short-wave infrared (SWIR) camera
AI summary Read the abstract
Researchers demonstrated that a lightweight hyperspectral short-wave infrared camera mounted on a drone could successfully detect and distinguish polyethylene and polypropylene plastics on beaches under natural sunlight, establishing proof of concept for aerial plastic monitoring.
A Deep Learning Model for Automatic Plastic Mapping Using Unmanned Aerial Vehicle (UAV) Data
AI summary Read the abstract
Researchers applied a deep learning semantic segmentation model (ResUNet50 based on U-Net architecture) to UAV orthophotos to automatically map floating plastic debris, achieving F1-scores of 0.86-0.92 for specific plastic types including oriented polystyrene, nylon, and PET. Classification accuracy decreased with lower spatial resolution, with 4 mm resolution providing optimal performance for distinguishing plastic types.
Research digests by email
When a large batch of papers lands in the Atlas, we read through it and send a short write-up of what stood out.