Lin Li

Permanent URI for this collection

https://hdl.handle.net/1805/25251

Remote sensing of inland water quality

Outbreaks of cyanobacteria in inland water bodies, especially water sources, can pose a serious threat to public health. Studies have shown that public exposure or ingestion of cyanobacterial cells and toxins can cause a series of harmful health consequences, such as skin irritation, allergic reactions, mucous membrane blistering, muscle and joint pain, gastroenteritis, lung consolidation, liver and kidney damage, and various neurological effects. Furthermore, the outbreak of cyanobacteria can cause factory shutdowns and tremendous economic losses. Therefore, the scientific value and socio-economic significance of inland water quality remote sensing monitoring and early warning research are beyond doubt. Chlorophyll-a and phycocyanin are important indicators for monitoring and early warning of outbreaks of cyanobacteria in inland water bodies with remote sensing because both pigments exhibit diagnostic absorption spectral features in the visible spectral region.

Over the last 15 years, Dr. Lin Li's research group has been making efforts to improve remote sensing approaches to monitoring inland water quality. Particularly, the effort was put in deriving the absorption and scattering coefficients of optically active substances (OACs) in water from remote sensing reflectance spectra and decomposing the total absorption of OACs to achieve the separation of the absorption coefficients of algal pigments from other constituents such as soluble organic matter and suspended sediment. An accurate retrieval of pigment absorption coefficients makes it possible to reliably map chlorophyll-a and phycocyanin and assess inland water quality. This would provide water management authorities and managers with an effective remote sensing tool for dealing with the outbreak of cyanobacteria in inland water bodies.

Dr. Li's work to improve remote sensing and monitoring of inland water quality is another excellent example of how IUPUI's faculty members are TRANSLATING their RESEARCH INTO PRACTICE.

Browse

Now showing 1 - 10 of 29

Advancing cyanobacteria biomass estimation from hyperspectral observations: Demonstrations with HICO and PRISMA imagery
(Elsevier, 2021-12) O'Shea, Ryan E.; Pahlevan, Nima; Smith, Brandon; Bresciani, Mariano; Egerton, Todd; Giardino, Claudia; Li, Lin; Moore, Tim; Ruiz-Verdu, Antonio; Ruberg, Steve; Simis, Stefan G. H.; Stumpf, Richard; Vaičiūtė, Diana; Earth Sciences, School of Science
Retrieval of the phycocyanin concentration (PC), a characteristic pigment of, and proxy for, cyanobacteria biomass, from hyperspectral satellite remote sensing measurements is challenging due to uncertainties in the remote sensing reflectance (∆Rrs) resulting from atmospheric correction and instrument radiometric noise. Although several individual algorithms have been proven to capture local variations in cyanobacteria biomass in specific regions, their performance has not been assessed on hyperspectral images from satellite sensors. Our work leverages a machine-learning model, Mixture Density Networks (MDNs), trained on a large (N = 939) dataset of collocated in situ chlorophyll-a concentrations (Chla), PCs, and remote sensing reflectance (Rrs) measurements to estimate PC from all relevant spectral bands. The performance of the developed model is demonstrated via PC maps produced from select images of the Hyperspectral Imager for the Coastal Ocean (HICO) and Italian Space Agency's PRecursore IperSpettrale della Missione Applicativa (PRISMA) using a matchup dataset. As input to the MDN, we incorporate a combination of widely used band ratios (BRs) and line heights (LHs) taken from existing multispectral algorithms, that have been proven for both Chla and PC estimation, as well as novel BRs and LHs to increase the overall cyanobacteria biomass estimation accuracy and reduce the sensitivity to ∆Rrs. When trained on a random half of the dataset, the MDN achieves uncertainties of 44.3%, which is less than half of the uncertainties of all viable optimized multispectral PC algorithms. The MDN is notably better than multispectral algorithms at preventing overestimation on low (<10 mg m−3) PC. Visibly, HICO and PRISMA PC maps show the wider dynamic range that can be represented by the MDN. The available in situ and satellite-derived Rrs matchups and measured in situ PC demonstrate the robustness of the MDN for estimating low (<10 mg m−3) PC and the reduced impact of ∆Rrs on medium-to-high in situ PC (>10 mg m−3). According to our extensive assessments, the developed model is anticipated to enable practical PC products from PRISMA and HICO, therefore the model is promising for planned hyperspectral missions, such as the Plankton Aerosol and Cloud Ecosystem (PACE). This advancement will enhance the complementary roles of hyperspectral radiometry from satellite and low-altitude platforms for quantifying and monitoring cyanobacteria harmful algal blooms at both large and local spatial scales.
BAND SELECTION METHOD APPLIED TO M3 (MOON MINERALOGY MAPPER)
(Office of the Vice Chancellor for Research, 2012-04-13) Cavanagh, Patrick D.; Li, Lin
Remote sensing optical sensors, such as those on board satellites and planetary probes, are able to detect and measure solar radiation at both im-proved spectral and spatial resolution. In particular, a hyperspectral dataset often consists of tens to hundreds of specified wavelength bands and con-tains a vast amount of spectral information for potential processing. One drawback of such a large spectral dataset is information redundancy result-ing from high correlation between narrow spectral bands. Reducing the data dimensionality is critical in practical hyperspectral remote sensing applica-tions. Price’s method is a band selection approach that uses a small subset of bands to accurately reconstruct the full hyperspectral dataset. The method seeks to represent the dataset by a weighted sum of basis functions. An it-erative process is used to successively approximate the full dataset. The process ends when the last basis function no longer provides a significant contribution to the reconstruction of the dataset, i.e. the basis function is dominated by noise. The research presented examines the feasibility of Price’s method for ex-tracting an optimal band subset from recently acquired lunar hyperspectral images recorded by the Moon Mineralogy Mapper (M3) instrument on board the Chandrayaan-1 spacecraft. The Apollo 17 landing site was used for test-ing of the band selection method. Preliminary results indicate that the band selection method is able to successfully reconstruct the original hyperspectral dataset with minimal error. In a recent test case, 15 bands were used to reconstruct the original 74 bands of reflectance data. This represents an accurate reconstruction using only 20% of the original dataset. The results from this study can help to configure spectral channels of fu-ture optical instruments for lunar exploration. The channels can be chosen based on the knowledge of which wavelength bands represent the greatest relevant information for characterizing geology of a particular location.
A Comparative Assessment of Geostatistical, Machine Learning, and Hybrid Approaches for Mapping Topsoil Organic Carbon Content
(MDPI, 2019-04) Chen, Lin; Ren, Chunying; Li, Lin; Wang, Yeqiao; Zhang, Bai; Wang, Zongming; Li, Linfeng; Earth Sciences, School of Science
Accurate digital soil mapping (DSM) of soil organic carbon (SOC) is still a challenging subject because of its spatial variability and dependency. This study is aimed at comparing six typical methods in three types of DSM techniques for SOC mapping in an area surrounding Changchun in Northeast China. The methods include ordinary kriging (OK) and geographically weighted regression (GWR) from geostatistics, support vector machines for regression (SVR) and artificial neural networks (ANN) from machine learning, and geographically weighted regression kriging (GWRK) and artificial neural networks kriging (ANNK) from hybrid approaches. The hybrid approaches, in particular, integrated the GWR from geostatistics and ANN from machine learning with the estimation of residuals by ordinary kriging, respectively. Environmental variables, including soil properties, climatic, topographic, and remote sensing data, were used for modeling. The mapping results of SOC content from different models were validated by independent testing data based on values of the mean error, root mean squared error and coefficient of determination. The prediction maps depicted spatial variation and patterns of SOC content of the study area. The results showed the accuracy ranking of the compared methods in decreasing order was ANNK, SVR, ANN, GWRK, OK, and GWR. Two-step hybrid approaches performed better than the corresponding individual models, and non-linear models performed better than the linear models. When considering the uncertainty and efficiency, ML and two-step approach are more suitable than geostatistics in regional landscapes with the high heterogeneity. The study concludes that ANNK is a promising approach for mapping SOC content at a local scale.
Convolutional neural network model for soil moisture prediction and its transferability analysis based on laboratory Vis-NIR spectral data
(Elsevier, 2021-12) Chen, Yu; Li, Lin; Whiting, Michael; Chen, Fang; Sun, Zhongchang; Song, Kaishan; Wang, Qinjun; Earth Sciences, School of Science
Laboratory visible near infrared reflectance (Vis-NIR, 400–2500 nm) spectroscopy has the advantages of simplicity, fast and non-destructive which was used for SM prediction. However, many previously proposed models are difficult to transfer to unknown target areas without recalibration. In this study, we first developed a suitable Convolutional Neutral Network (CNN) model and transferred the model to other target areas for two situations using different soil sample backgrounds under 1) the same measurement conditions (DSSM), and 2) under different measurement conditions (DSDM). We also developed the CNN models for the target areas based on their own datasets and traditional PLS models was developed to compare their performances. The results show that one dimensional model (1D-CNN) performed strongly for SM prediction with average R2 up to 0.989 and RPIQ up to 19.59 in the laboratory environment (DSSM). Applying the knowledge-based transfer learning method to an unknown target area improved the R2 from 0.845 to 0.983 under the DSSM and from 0.298 to 0.620 under the DSDM, which performed better than data-based spiking calibration method for traditional PLS models. The results show that knowledge-based transfer learning was suitable for SM prediction under different soil background and measurement conditions and can be a promising approach for remotely estimating SM with the increasing amount of soil dataset in the future.
Coupling Coordination Relationship between Urban Sprawl and Urbanization Quality in the West Taiwan Strait Urban Agglomeration, China: Observation and Analysis from DMSP/OLS Nighttime Light Imagery and Panel Data
(MDPI, 2020-10) Lu, Chunyan; Li, Lin; Lei, Yifan; Ren, Chunying; Su, Ying; Huang, Yufei; Chen, Yu; Lei, Shaohua; Fu, Weiwei; Earth Sciences, School of Science
Urban sprawl is the most prominent characteristic of urbanization, and increasingly affects local and regional sustainable development. The observation and analysis of urban sprawl dynamics and their relationship with urbanization quality are essential for framing integrative urban planning. In this study, the urban areas of the West Taiwan Strait Urban Agglomeration (WTSUA) were extracted using nighttime light imagery from 1992 to 2013. The spatio-temporal characteristics and pattern of urban sprawl were quantitatively analyzed by combining an urban expansion rate index and a standard deviation ellipse model. The urbanization quality was assessed using an entropy weight model, and its relationship with urban sprawl was calculated by a coupling coordination degree model. The results showed that the urban area in the WTSUA experienced a significant increase, i.e., 18,806.73 km2, during the period 1992–2013. The central cities grew by 11.08% and noncentral cities by 27.43%, with a general uneven city rank-size distribution. The urban sprawl showed a circular expansion pattern, accompanied by a gradual centroid migration of urban areas from the southeast coast to the central-western regions. The coupling coordination level between urban expansion and urbanization quality increased from serious incoordination in 1992 to basic coordination in 2013. Dual driving forces involving state-led policies and market-oriented land reform had a positive influence on the harmonious development of urban sprawl and urbanization quality of the WTSUA. This research offers an effective approach to monitor changes in urban sprawl and explore the coupling coordination relationship between urban sprawl and urbanization quality. The study provides important scientific references for the formulation of future policies and planning for sustainable development in urban agglomerations.
Cyanobacteria in Inland Waters: Remote Sensing
(CRC Press, 2020) Li, Lin; Earth Sciences, School of Science
Remote sensing plays important roles in managing harmful cyanobacterial blooms. Remote sensing algorithms for monitoring cyanobacterial blooms are grouped into empirical, semi-empirical, and semi-analytical methods. In this chapter, 12 of these methods were selected to be reviewed for their performances when applied to in situ measured field reflectance spectra and airborne or satellite sensor collected image spectra. Five empirical PC algorithms based on either band ratio or baseline calculation showed data-dependent performances, empirical band ratios and the baseline can be used to build semi-empirical models such as double three band baseline (DTBB) and four band baseline model (FBBM) showing stronger performance than the three band model (TBM), and the DTBB even performing stronger than the nested band ratio (NBR). As far as three semi-analytical models concern, the NBR and EIIMIW consistently performed well compared to the QAA pc , but care or recalibration should be practiced for applying both EIIMIW and NBR given caring inherent optical property of non-phycocyanin (PC) constituent in the water column. Although neither DTBB nor FBM cannot be evaluated with satellite MEdium Resolution Imaging Spectrometer (MERIS) and Ocean and Land Color Instrument (OLCI) images, they should be tested in future with hyperspectral satellite images acquired by PRISMA, EnMAP, and HyspIRI.
Determine characteristics requirement for the surrogate road edge objects for road departure mitigation testing
(2019) Chien, Stanley; Yi, Qiang; Lin, Jun; Saha, Abir; Li, Lin; Chen, Yaobin; Chen, Chi-Chih; Sherony, Rini; Electrical and Computer Engineering, School of Engineering and Technology
Road departure mitigation system (RDMS), a vehicle active safety feature, uses road edge objects to determine potential road departure. In the U.S., 45%, 16%, and 15% of car-mile (traffic flow * miles) roads have grass, metal guardrail, and concrete divider as road edge, respectively. It is difficult to test RDMS with real roadside objects. Lightweight and crashable surrogate roadside objects that have representative radar, LIDAR and camera characteristics of real objects have been developed for testing. This paper describes the identification of automotive radar, LIDAR, and visual characteristics of metal guardrail, concrete divider, and grass. These characteristics will be referenced for designing and fabricating the representative surrogate objects for RDMS testing. Colors and types of the roadside objects were identified from 24,735 randomly sampled locations in the US using Google street view images. The radar and LIDAR parameters were measured using 24GHz/77GHz radar and 350-2500nm IR spectrometer.
Hyperspectral retrievals of phytoplankton absorption and chlorophyll-a in inland and nearshore coastal waters
(Elsevier, 2021-02) Pahlevan, Nima; Smith, Brandon; Binding, Caren; Gurlin, Daniela; Li, Lin; Bresciani, Mariano; Giardino, Claudia; Earth Sciences, School of Science
Following more than two decades of research and developments made possible through various proof-of-concept hyperspectral remote sensing missions, it has been anticipated that hyperspectral imaging would enhance the accuracy of remotely sensed in-water products. This study investigates such expected improvements and demonstrates the utility of hyperspectral radiometric measurements for the retrieval of near-surface phytoplankton properties1, i.e., phytoplankton absorption spectra (aph) and biomass evaluated through examining the concentration of chlorophyll-a (Chla). Using hyperspectral data (409–800 nm at ~5 nm resolution) and a class of neural networks known as Mixture Density Networks (MDN) (Pahlevan et al., 2020), we show that the median error in aph retrievals is reduced two-to-three times (N = 722) compared to that from heritage ocean color algorithms. The median error associated with our aph retrieval across all the visible bands varies between 20 and 30%. Similarly, Chla retrievals exhibit significant improvements (i.e., more than two times; N = 1902), with respect to existing algorithms that rely on select spectral bands. Using an independent matchup dataset acquired near-concurrently with the acquisition of the Hyperspectral Imager for the Coastal Ocean (HICO) images, the models are found to perform well, but at reduced levels due to uncertainties in the atmospheric correction. The mapped spatial distribution of Chla maps and aph spectra for selected HICO swaths further solidify MDNs as promising machine-learning models that have the potential to generate highly accurate aquatic remote sensing products in inland and coastal waters. For aph retrieval to improve further, two immediate research avenues are recommended: a) the network architecture requires additional optimization to enable a simultaneous retrieval of multiple in-water parameters (e.g., aph, Chla, absorption by colored dissolved organic matter), and b) the training dataset should be extended to enhance model generalizability. This feasibility analysis using MDNs provides strong evidence that high-quality, global hyperspectral data will open new pathways toward a better understanding of biodiversity in aquatic ecosystems.
The impact of fog on soil moisture dynamics in the Namib Desert
(Elsevier, 2018-03) Li, Bonan; Wang, Lixin; Kaseke, Kudzai F.; Vogt, Roland; Li, Lin; Seely, Mary; Earth Science, School of Science
Soil moisture is a crucial component supporting vegetation dynamics in drylands. Despite increasing attention on fog in dryland ecosystems, the statistical characterization of fog distribution and how fog affects soil moisture dynamics have not been seen in literature. To this end, daily fog records over two years (Dec 1, 2014–Nov 1, 2016) from three sites within the Namib Desert were used to characterize fog distribution. Two sites were located within the Gobabeb Research and Training Center vicinity, the gravel plains and the sand dunes. The third site was located at the gravel plains, Kleinberg. A subset of the fog data during rainless period was used to investigate the effect of fog on soil moisture. A stochastic modeling framework was used to simulate the effect of fog on soil moisture dynamics. Our results showed that fog distribution can be characterized by a Poisson process with two parameters (arrival rate λ and average depth α (mm)). Fog and soil moisture observations from eighty (Aug 19, 2015–Nov 6, 2015) rainless days indicated a moderate positive relationship between soil moisture and fog in the Gobabeb gravel plains, a weaker relationship in the Gobabeb sand dunes while no relationship was observed at the Kleinberg site. The modeling results suggested that mean and major peaks of soil moisture dynamics can be captured by the fog modeling. Our field observations demonstrated the effects of fog on soil moisture dynamics during rainless periods at some locations, which has important implications on soil biogeochemical processes. The statistical characterization and modeling of fog distribution are of great value to predict fog distribution and investigate the effects of potential changes in fog distribution on soil moisture dynamics.
The Impact of Rainfall on Soil Moisture Dynamics in a Foggy Desert.
(PLOS, 2016) Li, Bonan; Wang, Lixin; Kaseke, Kudzai F.; Li, Lin; Seely, Mary K.; Department of Earth Sciences, School of Science
Soil moisture is a key variable in dryland ecosystems since it determines the occurrence and duration of vegetation water stress and affects the development of weather patterns including rainfall. However, the lack of ground observations of soil moisture and rainfall dynamics in many drylands has long been a major obstacle in understanding ecohydrological processes in these ecosystems. It is also uncertain to what extent rainfall controls soil moisture dynamics in fog dominated dryland systems. To this end, in this study, twelve to nineteen months’ continuous daily records of rainfall and soil moisture (from January 2014 to August 2015) obtained from three sites (one sand dune site and two gravel plain sites) in the Namib Desert are reported. A process-based model simulating the stochastic soil moisture dynamics in water-limited systems was used to study the relationships between soil moisture and rainfall dynamics. Model sensitivity in response to different soil and vegetation parameters under diverse soil textures was also investigated. Our field observations showed that surface soil moisture dynamics generally follow rainfall patterns at the two gravel plain sites, whereas soil moisture dynamics in the sand dune site did not show a significant relationship with rainfall pattern. The modeling results suggested that most of the soil moisture dynamics can be simulated except the daily fluctuations, which may require a modification of the model structure to include non-rainfall components. Sensitivity analyses suggested that soil hygroscopic point (sh) and field capacity (sfc) were two main parameters controlling soil moisture output, though permanent wilting point (sw) was also very sensitive under the parameter setting of sand dune (Gobabeb) and gravel plain (Kleinberg). Overall, the modeling results were not sensitive to the parameters in non-bounded group (e.g., soil hydraulic conductivity (Ks) and soil porosity (n)). Field observations, stochastic modeling results as well as sensitivity analyses provide soil moisture baseline information for future monitoring and the prediction of soil moisture patterns in the Namib Desert.

Browse

Browsing Lin Li by Title

Results Per Page

Sort Options