scholarly journals Inland Waters Suspended Solids Concentration Retrieval Based on PSO-LSSVM for UAV-Borne Hyperspectral Remote Sensing Imagery

2019 ◽  
Vol 11 (12) ◽  
pp. 1455 ◽  
Author(s):  
Lifei Wei ◽  
Can Huang ◽  
Yanfei Zhong ◽  
Zhou Wang ◽  
Xin Hu ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Least Squares ◽  
Empirical Model ◽  
Suspended Solids ◽  
Hyperspectral Remote Sensing ◽  
Support Vector ◽  
Inland Waters ◽  
Important Indicator ◽  
Solids Concentration ◽  
Semi Empirical

Suspended solids concentration (SSC) is an important indicator of the degree of water pollution. However, when using an empirical or semi-empirical model adapted to some of the inland waters to estimate SSC on unmanned aerial vehicle (UAV)-borne hyperspectral images, the accuracy is often not sufficient. Thus, in this study, we attempted to use the particle swarm optimization (PSO) algorithm to find the optimal parameters of the least-squares support vector machine (LSSVM) model for the quantitative inversion of SSC. A reservoir and a polluted riverway were selected as the study areas. The spectral data of the 36-point and 29-point 400–900 nm wavelength range on the UAV-borne images were extracted. Compared with the semi-empirical model, the random forest (RF) algorithm and the competitive adaptive reweighted sampling (CARS) algorithm combined with partial least squares (PLS), the accuracy of the PSO-LSSVM algorithm in predicting the SSC was significantly improved. The training samples had a coefficient of determination ( R 2 ) of 0.98, a root mean square error (RMSE) of 0.68 mg/L, and a mean absolute percentage error (MAPE) of 12.66% at the reservoir. For the polluted riverway, PSO-LSSVM also performed well. Finally, the established SSC inversion model was applied to UAV-borne hyperspectral remote sensing (HRS) images. The results confirmed that the distribution of the predicted SSC was consistent with the observed results in the field, which proves that PSO-LSSVM is a feasible approach for the SSC inversion of UAV-borne HRS images.

Spectral reflectance of suspended solid sediments: a laboratory experiment.

2020 ◽  
Author(s):  
Martina Carlino ◽  
Silvia Di Francesco
Keyword(s):  
Remote Sensing ◽  
Laboratory Experiment ◽  
Spectral Reflectance ◽  
Suspended Solids ◽  
Ocean Color ◽  
Total Suspended Solids ◽  
Reflectance Spectra ◽  
Optical Data ◽  
Band Selection ◽  
Solids Concentration

<p>Ocean color remote sensing proved to be a good alternative to traditional methods for total suspended solids concentration (TSS) monitoring purposes: numerous sensors have been developed for ocean color applications and different algorithms to retrieve TSS from remotely sensed data already exist.</p><p>Nevertheless, their application is generally limited by site-specific factors, and presently there is no uniform remote sensing model to estimate TSS.</p><p>The present study is focused in the development, evaluation and validation of different algorithms to estimate total suspended solids concentration based on laboratory reflectance data.</p><p>At this aim, a laboratory experiment was designed to collect the spectral reflectance of water containing fixed suspended particulate matter in terms of its concentration.</p><p>During the experiment, a total of 10 silty clay loam sediment samples were introduced into a tank filled with clear water up to a depth of 22 cm, illuminated by two 45 W lamps focused on center of water surface. After sieving, sediments were weighed so that TSS concentration ranging from 150 up to 2000 mg/L were obtained in the tank, being soil sediments suspension guaranteed by means of a mechanical pump-driven device.</p><p>Optical data were collected few minutes after each sediment introduction, using an Ocean Optics Jaz spectroradiometer mounted on a platform above the tank.</p><p>In accordance with previous studies, collected reflectance spectra of water containing sediments showed that, whatever is sediment concentration in water, reflectance in the red region is larger than that in the NIR region. Furthermore, reflectance spectra generally present two peaks: one between 550 nm and 750 nm, and the other between 750 nm and 850 nm, being the second peak insignificant for samples with very small TSS (e.g., SSC=150 mg/L), due to strong absorption of water.</p><p>After collection, laboratory reflectance spectra were integrated over the bandpass of different sensors’ selected bands, modulated by their relative response functions (RSR).</p><p>The basic principle of using a specific band, or band ratios to estimate a water parameter is to select spectral bands representative of its scattering/absorption features.</p><p>Band selection was achieved testing some previously formulated ocean color algorithms for the estimation of water quality parameters.</p><p>After band selection, linear regression model was applied to estimate the relationship between sensors’ reflectance at these bands and suspended solids concentration.</p><p>Results showed high correlation between selected sensors’ spectral red band and total suspended solids concentration higher than 500 mg/L up to 1360 mg/L, while less accuracy was observed for TSS concentrations higher than 1360 mg/L. Furthermore, the ratio between spectral red and green bands better estimates TSS in waters where total suspended concentration is not higher than 500 mg/L.</p><p> </p>

scholarly journals Diversity of Remote Sensing-Based Variable Inputs Improves the Estimation of Seasonal Maximum Freezing Depth

2021 ◽  
Vol 13 (23) ◽  
pp. 4829
Author(s):  
Bingquan Wang ◽  
Youhua Ran
Keyword(s):  
Machine Learning ◽  
Remote Sensing ◽  
Thermal State ◽  
Reanalysis Data ◽  
Soil Freezing ◽  
Machine Learning Techniques ◽  
Tibet Plateau ◽  
Support Vector ◽  
Important Indicator ◽  
Freezing Depth

The maximum soil freezing depth (MSFD) is an important indicator of the thermal state of seasonally frozen ground. Its variation has important implications for the water cycle, ecological processes, climate and engineering stability. This study tested three aspects of data-driven predictions of MSFD in the Qinghai-Tibet Plateau (QTP), including comparison of three popular statistical/machine learning techniques, differences between remote sensing variables and reanalysis data as input conditions, and transportability of the model built by reanalysis data. The results show that support vector regression (SVR) performs better than random forest (RF), k-nearest neighbor (KNN) and the ensemble mean of the three models. Compared with the climate predictors, the remote sensing predictors are helpful for improving the simulation accuracy of the MSFD at both decadal and annual scales (at the annual and decadal scales, the root mean square error (RMSE) is reduced by 2.84 and 1.99 cm, respectively). The SVR model with climate predictor calibration using the in situ MSFD at the baseline period (2001–2010) can be used to simulate the MSFD over historical periods (1981–1990 and 1991–2000). This result indicates the good transferability of the well-trained machine learning model and its availability to simulate the MSFD of the past and the future when remote sensing predictors are not available.

Sign in / Sign up

Export Citation Format

Share Document