We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Multi-Scale Hierarchical Attention Ensemble Network for Fine-Grained Riverine Waste Segmentation Using UAV Multispectral Imagery
AI summary Read the abstract
Researchers used a drone with a special five-band camera to fly over a river in Indonesia and trained an AI system to automatically spot and map plastic waste floating in the water, even distinguishing it from natural debris like leaves and branches. This matters because rivers are a major pathway for plastic to reach oceans and break down into microplastics, which can end up in our food and water; better, faster detection tools like this could help communities target cleanup efforts before waste spreads further. The study is an early proof-of-concept from one river during one season, so more testing is needed before it can be used widely.
Riverine plastic waste is difficult to detect and map accurately because debris ranges from small items to large floating clusters, and tropical rivers present challenging conditions, such as murky water, floating vegetation, and variable lighting. This study develops and benchmarks a deep learning method for pixel-level, multi-scale mapping of riverine waste from five-band UAV multispectral imagery. We deployed a drone equipped with a five-band multispectral sensor over the Brantas River in Surabaya, East Java, Indonesia, and introduce the Multi-Scale Hierarchical Attention Ensemble (MHAE), which combines three backbone networks across three image resolutions through learned scale- and backbone-attention weighting. MHAE was evaluated against 12 CNN-, transformer-, state-space-, and traditional-machine-learning-based baselines (including Random Forest, U-Net, UNet++, and DeepLabV3+) on the accompanying BrantasRiverWaste-UAV dataset (882 image tiles from a single-site, single-season orthomosaic covering approximately 0.54 km2 of river surface, labelled as water, land, organic waste, or inorganic waste), with pairwise comparisons assessed using Wilcoxon signed-rank tests with Bonferroni correction. MHAE achieved the highest pixel-level waste detection rate among the 12 evaluated models (86.76%), with a mean intersection-over-union of 78.33% (third-highest, behind UNet++ and U-Net). This work provides an initial, single-site benchmark and reference architecture for near-real-time riverine waste monitoring, and introduces the BrantasRiverWaste-UAV as, to our knowledge, one of the first five-band multispectral UAV datasets for tropical riverine waste mapping. Balanced sampling and multi-scale attention fusion can substantially improve waste-pixel detection under severe class imbalance; because the benchmark derives from a single site and season, future work should extend evaluation across additional seasons and river systems before the approach is generalised operationally.
More Papers Like This
Real-Time Riverine Plastic Waste Detection Using an Enhanced YOLO Framework
AI summary Read the abstract
Researchers built an AI tool that can automatically spot plastic waste in rivers with about 95% accuracy, using smart camera technology similar to what's used in self-driving cars. This matters because plastic trash in rivers eventually breaks down into microplastics that can end up in our drinking water and food supply—catching it early and tracking pollution hotspots could help communities clean up waterways before that happens.
FCA-YOLOv8n: Frequency-Augmented Lightweight Network for River Plastic Waste Detection
AI summary Read the abstract
Researchers built a lightweight AI system that drones can use to spot plastic trash floating in rivers—even small, clear, or shiny pieces that are normally hard to see—while running efficiently enough for real-world use. This matters because catching plastic waste in rivers before it breaks down into microplastics could help reduce the amount that ends up in our water, food, and eventually our bodies, where microplastics have raised health concerns. While this study focuses on detection technology rather than health outcomes directly, better tracking tools are a key step toward cleaning up waterways more effectively.
Designing Realtime Object Detection Based Plastic Litter Monitoring System
AI summary Read the abstract
Researchers built an automated camera system that uses AI to detect and count plastic waste floating in rivers, offering a faster alternative to the current method in Indonesia, which relies on people manually watching and counting debris. This matters because rivers are a major pathway for plastic pollution into the ocean, and better tracking tools can help pinpoint pollution sources and guide cleanup efforts before the plastic breaks down into microplastics that contaminate water, food, and potentially our bodies. While the AI system slightly undercounted debris compared to human observers, it offers the advantage of continuous, real-time monitoring that isn't practical
Ocean and River Plastic Waste Detection Using Computer Vision
AI summary Read the abstract
Millions of tons of plastic pour into our oceans every year, much of it carried by rivers—but tracking where it comes from has always relied on slow, manual surveys. This study built an AI camera system that automatically spots and counts floating plastic waste from drone and shoreline footage with over 91% accuracy, offering a faster way to find pollution hotspots. Better tracking matters because plastic waste that lingers in waterways breaks down into microplastics, which end up in the fish and water we consume, making tools like this useful for guiding cleanup efforts before pollution reaches our food and water supply.
Develop an Automated system for River Water Pollutant Detection and Classification to Alert the Authority using a Deep Learning Framework
AI summary Read the abstract
Scientists built an AI-powered camera system that can automatically spot trash and pollution—like plastic, oil, and foam—floating in rivers, then instantly alert cleanup crews to the exact location. This matters because polluted rivers often feed into drinking water supplies and can carry microplastics and toxins that harm human health, so catching pollution faster means it can be cleaned up before it spreads further into the water we rely on.
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.