Semantic segmentation of UAV remote sensing images based on edge feature fusing and multi-level upsampling integrated with Deeplabv3.

Deeplabv3+ currently is the most representative semantic segmentation model. However, Deeplabv3+ tends to ignore targets of small size and usually fails to identify precise segmentation boundaries in the UAV remote sensing image segmentation task. To handle these problems, this paper proposes a semantic segmentation algorithm of UAV remote sensing images based on edge feature fusing and multi-level upsampling integrated with Deeplabv3+ (EMNet). EMNet uses MobileNetV2 as its backbone and adds an edge detection branch in the encoder to provide edge information for semantic segmentation. In the decoder, a multi-level upsampling method is designed to retain high-level semantic information (e.g., the target's location and boundary information). The experimental results show that the mIoU and mPA of EMNet improved over Deeplabv3+ by 7.11% and 6.93% on the dataset UAVid, and by 0.52% and 0.22% on the dataset ISPRS Vaihingen.

A UAV and S2A data-based estimation of the initial biomass of green algae in the South Yellow Sea.

Previous studies have shown that the initial biomass of green tide was the green algae attaching to Pyropia aquaculture rafts in the Southern Yellow Sea. In this study, the green algae was identified with unmanned aerial vehicle (UAV), an biomass estimation model was proposed for green algae biomass in the radial sand ridge area based on Sentinel-2A image (S2A) and UAV images. The result showed that the green algae was detected highly accurately with the normalized green-red difference index (NGRDI); approximately 1340 tons and 700 tons of green algae were attached to rafts and raft ropes respectively, and the lower biomass might be the main cause for the smaller scale of green tide in 2017. In addition, UAV play an important role in raft-attaching green algae monitoring and long-term research of its biomass would provide a scientific basis for the control and forecast of green tide in the Yellow Sea.

GIS-based health assessment of the marine ecosystem in Laizhou Bay, China.

According to 2014-2016 monitoring data, an assessment index system including water quality, depositional environment and ecosystem was built to evaluate the health statue of marine ecosystem in the Laizhou Bay using analytic hierarchy process (AHP) method. The results, spatialized in ArcGIS software, show: while the comprehensive ecological health index is 0.62, the ecological environmental quality in the Laizhou Bay is in a sub-healthy state; the unhealthy area is mainly concentrated in southwestern inshore region, and impacted by serious environmental problems, such as water eutrophication and heavy metal pollution; the northwestern and southeastern inshore regions are in a sub-healthy state, while the eastern inshore and northern areas are in the healthiest state. The land-based pollutants that discharge into the sea may be the leading factors that are causing ecological environment deterioration in the Laizhou Bay, and the reclamation work ongoing around the port has exacerbated the ecological risk.

[Methane fluxes of Cyperus malaccensis tidal wetland in Minjiang River estuary].

By using enclosed static chamber-gas chromatograph techniques, this paper measured the methane fluxes of Cyperus malaccensis tidal wetland in Minjiang River estuary. The diurnal variation of the methane fluxes in summer and winter were in the range of 1.29-2.93 mg x m(-2) x h(-1) and 0.06-0.22 mg x m(-2) x h(-1), respectively. The methane fluxes before flooding, in the process of flooding and ebbing, and after ebbing were 0.11-1.52 mg x m(-2) x h(-1), 0.10-1.05 mg x m(-2) x h(-1), and 0.05-1.70 mg x m(-2) x h(-1), and the monthly averaged fluxes were 0.73, 0.47, and 0.72 mg x m(-2) x h(-1), respectively. The methane fluxes peaked in September and reached the lowest in March, and were significantly lower in the process of flooding and ebbing than before flooding and after ebbing (P < 0.05). The seasonal variation of the methane fluxes was in the order of summer > autumn > spring > winter. Tide was the key factor affecting the diurnal variation of the methane fluxes, while plant growth stage and temperature were the key factors determining the monthly or seasonal variation of the methane fluxes.

[Seasonal dynamics of nitrogen- and phosphorus absorption efficiency of wetland plants in Minjiang River estuary].

Taking the native Phragmites australis and invasive Spartina alterniflora in Minjiang River estuary as test objectives, this paper studied the seasonal dynamics of their biomass and nitrogen- and phosphorus absorption efficiency. A typical single-peak curve was presented for the seasonal dynamics of aboveground biomass and nitrogen- and phosphorus absorption efficiency of the two species. P. australis had the maximum aboveground biomass (2195.33 g X m(-2)) in summer, while S. alterniflora had it (3670.02 g X m(-2)) in autumn. The total nitrogen (TN) and total phosphorus (TP) contents of P. australis reached the peak (21.06 g x m(-2) of TN and 1.12 g x m(-2) of TP) in summer and in autumn, respectively, while those of S. alterniflora all reached the peak (26.76 g x m(-2) of TN and 3.23 g x m(-2) of TP) in autumn. Both of the two species had a higher absorption efficiency in TN than in TP (P < 0.01), and S. alterniflora had a significantly higher absorption efficiency of TN and TP than P. australis (P < 0.05). To some extent, the N/P, C/N, and C/P ratios of plants could indicate the nitrogen- and phosphorus absorption efficiency of the plants.