Combining Deep Learning with Single-Spectrum UV Imaging for Rapid Detection of HNSs Spills

Syed Raza Mehdi,Amjad Ali,Hong Song,Kazim Raza,Rizwan Ali Naqvi,Hui Huang

doi:10.3390/rs14030576

Abstract

Vital transportation of hazardous and noxious substances (HNSs) by sea occasionally suffers spill incidents causing perilous mutilations to off-shore and on-shore ecology. Consequently, it is essential to monitor the spilled HNSs rapidly and mitigate the damages in time. Focusing on on-site and early processing, this paper explores the potential of deep learning and single-spectrum ultraviolet imaging (UV) for detecting HNSs spills. Images of three floating HNSs, including benzene, xylene, and palm oil, captured in different natural and artificial aquatic sites were collected. The image dataset involved UV (at 365 nm) and RGB images for training and comparative analysis of the detection system. The You Only Look Once (YOLOv3) deep learning model is modified to balance the higher accuracy and swift detection. With the MobileNetv2 backbone architecture and generalized intersection over union (GIoU) loss function, the model achieved mean IoU values of 86.57% for UV and 82.43% for RGB images. The model yielded a mean average precision (mAP) of 86.89% and 72.40% for UV and RGB images, respectively. The average speed of 57 frames per second (fps) and average detection time of 0.0119 s per image validated the swift performance of the proposed model. The modified deep learning model combined with UV imaging is considered computationally cost-effective resulting in precise detection accuracy and significantly faster detection speed.

Highlights

According to the International Maritime Organization, different chemical substances, such as petrochemical products and vegetable oil other than crude oil, which varies in physical and chemical properties [1], are considered colorless, hazardous, and noxious substances (HNSs)
The average IoU values of detection explain that the bounding box regression loss function based on generalized intersection over union (GIoU) better fits the proposed YOLOv3 lightweight model
The results conclude that the detection rate of palm oil has significantly better IoU values in both ultraviolet imaging (UV) and RGB images

Summary

Introduction

According to the International Maritime Organization, different chemical substances, such as petrochemical products and vegetable oil other than crude oil, which varies in physical and chemical properties [1], are considered colorless, hazardous, and noxious substances (HNSs). Petrochemical products, such as benzene and xylene [2] exhibit a broader range of properties (e.g., dissolving, floating, sinking, evaporation, etc.) and different toxicity levels. These chemicals have both acute and long-term ecological effects and cannot be recoverable if spilled in the sea [3]. HNSs spills are considered one of the major causes of marine pollution, damaging aquatic and onshore human life and likely interfering with other legitimate uses of marine resources [6,7].

Methods

Results

Discussion

Conclusion