Automatic detection of suspected sewage discharge from coastal outfalls based on Sentinel-2 imagery.

Terrestrial pollution has a great impact on the coastal ecological environment, and widely distributed coastal outfalls act as the final gate through which pollutants flow into rivers and oceans. Thus, effectively monitoring the water quality of coastal outfalls is the key to protecting the ecological environment. Satellite remote sensing provides an attractive way to monitor sewage discharge. Selecting the coastal areas of Zhejiang Province, China, as an example, this study proposes an innovative method for automatically detecting suspected sewage discharge from coastal outfalls based on high spatial resolution satellite imageries from Sentinel-2. According to the accumulated in situ observations, we established a training dataset of water spectra covering various optical water types from satellite-retrieved remote sensing reflectance (Rrs). Based on the clustering results from unsupervised classification and different spectral indices, a random forest (RF) classification model was established for the optical water type classification and detection of suspected sewage. The final classification covers 14 optical water types, with type 12 and type 14 corresponding to the high eutrophication water type and suspected sewage water type, respectively. The classification result of model training datasets exhibited high accuracy with only one misclassified sample. This model was evaluated by historical sewage discharge events that were verified by on-site observations and demonstrated that it could successfully recognize sewage discharge from coastal outfalls. In addition, this model has been operationally applied to automatically detect suspected sewage discharge in the coastal area of Zhejiang Province, China, and shows broad application value for coastal pollution supervision, management, and source analysis.

Construction and Application of YOLOv3-Based Diatom Identification Model of Scanning Electron Microscope Images.

OBJECTIVES: To construct a YOLOv3-based model for diatom identification in scanning electron microscope images, explore the application performance in practical cases and discuss the advantages of this model. METHODS: A total of 25 000 scanning electron microscopy images were collected at 1 500× as an initial image set, and input into the YOLOv3 network to train the identification model after experts' annotation and image processing. Diatom scanning electron microscopy images of lung, liver and kidney tissues taken from 8 drowning cases were identified by this model under the threshold of 0.4, 0.6 and 0.8 respectively, and were also identified by experts manually. The application performance of this model was evaluated through the recognition speed, recall rate and precision rate. RESULTS: The mean average precision of the model in the validation set and test set was 94.8% and 94.3%, respectively, and the average recall rate was 81.2% and 81.5%, respectively. The recognition speed of the model is more than 9 times faster than that of manual recognition. Under the threshold of 0.4, the mean recall rate and precision rate of diatoms in lung tissues were 89.6% and 87.8%, respectively. The overall recall rate in liver and kidney tissues was 100% and the precision rate was less than 5%. As the threshold increased, the recall rate in all tissues decreased and the precision rate increased. The F1 score of the model in lung tissues decreased with the increase of threshold, while the F1 score in liver and kidney tissues with the increase of threshold. CONCLUSIONS: The YOLOv3-based diatom electron microscope images automatic identification model works at a rapid speed and shows high recall rates in all tissues and high precision rates in lung tissues under an appropriate threshold. The identification model greatly reduces the workload of manual recognition, and has a good application prospect.

A simulation study of mercury immobilization in estuary sediment microcosm by activated carbon/clay-based thin-layer capping under artificial flow and turbation.

In-situ thin layer capping (TLC) is a promising sediment remediation approach that has been shown effective in immobilizing contaminants from releasing to natural biotas and human beings. This research intended to comprehend the effectiveness of Hg immobilization by TLC under turbation condition via a microcosm study. Three TLC caps with different activated carbon (AC)/clay combinations were applied to actual Hg-contaminated estuary sediment (76.0 ± 2.6 mg-Hg/kg). The caps with AC (3%) + bentonite (3%) and AC (3%) + kaolin (3%) were efficient in reducing both total mercury (THg) and methylmercury (MeHg) concentrations in overlying water by 75-95% and 64-98%, respectively, in the later stage of 75-d operation. In contrast, the AC (3%) + montmorillonite (3%) cap did not show a significant reduction on THg and MeHg in the overlying water, probably due to the unstable, suspension property of montmorillonite. The stable caps showed higher resistance to Hg breakthrough under occasional turbation events; however, a labile cap appeared to have dramatic Hg breakthrough when turbation occurred. It is therefore essential to note that with unstable caps, turbation events may result in unwanted secondary resuspension of contaminants.