Mapping urban land cover based on spatial-spectral classification of hyperspectral remote-sensing data

In this article, an innovative classification framework for hyperspectral image data, based on both spectral and spatial information, is proposed. The main objective of this method is to improve the accuracy and efficiency of high-resolution land-cover mapping in urban areas. The spatial information is obtained by an enhanced marker-based minimum spanning forest (MMSF) algorithm. A pixel-based support vector machine (SVM) algorithm is first used to classify the hyperspectral image data, then the enhanced MMSF algorithm is applied in order to increase the accuracy of less accurately classified land-cover types. The enhanced MMSF algorithm is used as a binary classifier. These two classes are the low-accuracy class and remaining classes. Finally, the SVM algorithm is trained for classes with acceptable accuracy. In the proposed approach, namely MSF-SVM, the markers are extracted from the classification maps obtained by both SVM and watershed segmentation algorithms, and are then used to build the MSF. Three benchmark hyperspectral data sets are used for the assessment: Berlin, Washington DC Mall, and Quebec City. Experimental results demonstrate the superiority of the proposed approach compared with SVM and the original MMSF algorithms. It achieves approximately 5, 6, and 7% higher rates in kappa coefficients of agreement in comparison with the original MMSF algorithm for the Berlin, Washington DC Mall, and Quebec City data sets, respectively.     

Spatiotemporal variability of snow cover on Tibet,China using MODIS remote-sensing data

ABSTRACT

Snow cover is an important component of the cryosphere, and the study on spatial and temporal variations of snow cover is essential for understanding the consequences and impacts of climate change and water resources management. In this study, the temporal variation of snow-covered area (SCA) and spatial variability of snow-cover frequency (SCF) on Tibet is analysed based on the Moderate Resolution Imaging Spectroradiometer (MODIS)/Terra snow cover product (MOD10A2) from 2000 to 2015, and relationships with main climate variables are investigated. Results are as follows: (1) there is a very weak decreasing trend in annual mean SCA, and a slight increasing trend in autumn and winter and a slight decreasing trend in spring and more robust decreasing trend in summer for SCA are found. (2) The temporal variation of SCA is negatively correlated with temperature, whereas it is little correlated with corresponding precipitation. (3) The general trend of spatial SCF variation on Tibet, predominated by snow-cover variations in spring and autumn, tends to decrease in spring while it tends to increase in autumn. (4) The spatial variability of SCF is attributed to snow-cover variations in autumn and spring, which is more obvious in higher latitudes in autumn while it is more noticeable in lower-latitude southeastern plateau in spring. (5) The regions with higher variability of snow cover are main pastoral land and more prone to snow-related disaster in Tibet, becoming key zone of snow-cover monitoring and disaster prevention and mitigation.     

A modified model of surface temperature inversion based on Landsat 8 remote-sensing data and measured data

ABSTRACT

This article verified the error between inversed land surface temperature (LST) and measured LST and developed the modified model based on Landsat 8 remote-sensing data. First, a single-channel algorithm was used to invert the surface temperature using four Landsat 8 remote-sensing images and the LST of the 98 measured points were obtained meantime. Then, the modified model between inversed LST and measured LST was developed based on LST for the 74 measured points. Finally, the developed models were used to modify the inversion temperatures at other 24 measured points, and the mean absolute error (MAE) and mean square error (MSE) between the measured temperature and the inversed temperature before and after the modification were compared to verify the validity of the model. The results showed that the MAE and the MSE of temperature for the 24 measured points used for verification reduced by 0.26 and 0.20 K, respectively, after modification. The development of the modified model can provide an important reference for using Landsat 8 remote-sensing image to invert surface temperature in other regions.     

Identification of suspended sediment in coastal waters using airborne thematic mapper data

Remotely sensed data were collected using an airborne Daedalus thematic mapper, flown at 4000 m, over Swansea Bay, northern Bristol Channel, U.K. The area is one of high tidal range and wave activity. Fifty-eight surface-water/sediment samples were collected from three vessels as concurrent ‘sea-truth’ data. Samples were analysed for suspended sediment concentration (SSC)

Regression equations are developed for SSC using data in the 450-520 nm (channel 2), 520-600 nm (channel 3), 605-625 nm (channel 4) and 630-690 nm (channel 5) spectral bands

Localized flow patterns are visible in the imagery and are identified using suspended material as the passive tracer. Such patterns cannot be identified in temporally averaged data collected from research vessels

It is concluded that airborne remote sensing can contribute to the understanding of the movement of water and sediments on continental shelves, when used in association with conventional oceanographic sampling techniques.     

Parameterizing the aerodynamic roughness length on a regional scale based on multi-source remote-sensing data

Aerodynamic roughness length (z₀) is one of those important biophysical parameters that influence energy exchange at the land–atmosphere interface, so it is significant to quantify the z₀ accurately. In this article, a scheme parameterizing land-surface z₀ at regional scale has been approached based on multi-resource remote-sensing data, including lidar and optical remote sensing. First, we retrieved the regional vegetation height from lidar data of Geoscience Laser Altimeter System (GLAS) on board the Ice, Cloud, and land Elevation Satellite (ICESat), and then the z₀ values of vegetated land surface were calculated using height data and canopy area index retrieved from remote-sensing data. Finally, the wall-to-wall map of z₀ in January and July 2008 were developed. The conclusions are as follows. (1) The vertical and horizontal structures of vegetation can be retrieved combining spaceborne lidar data and other optical remote-sensing data, so the vegetation characteristics and their intra-annual diversification of different land surfaces can be presented dynamically. The variation of z₀ with vegetation phenology can be quantified by modelling with vegetation height and multi-temporal leaf area index from multi-resource remote-sensing data. (2) The z₀ values of vegetated surface change significantly during leaf-on or leaf-off period in the year, but there are different features in the sparsely or densely vegetated surface. In the sparse vegetation areas, due to the relatively low leaf density in leaf-off season, the value of z₀ is also low. With the increase of leaf density in leaf-on season, the z₀ values will also increase. However, the relationship is complicated in the dense vegetation areas in leaf-on season; the z₀ values may or may not increase, but the zero-plane displacement heights will keep increasing continuously. This operational scheme to parameterize z₀ based on the vegetation height and canopy area index retrieved from multi-source remote-sensing data can be applied to quantify time serial z₀ at regional scale. Besides, it can also improve z₀ parameterization in land models or atmospheric models.     

Distribution and evolutionary of turfy soil identified by remote-sensing images based on fuzzy evaluation

Turfy soil is a kind of special soil accumulated by undecomposed plants which is detrimental to the engineering. In this paper, particular identified patterns for the turfy soil in the northeast of China was raised including the method of extracting threshold value, block analysis and fuzzy evaluation. And field investigations were undertaken to verify the accuracy of identification by remote sensing, and the correlation of field result and remote sensing result was summarized so as to analyze the regularities of distribution and evolutionary mechanism of turfy soil. The result shows that the combination of extracting threshold value, block analysis and fuzzy evaluation are effective methods to predict the distribution of the turfy soil; with the correlation of membership degree and field result, we can analyze the evolutionary mechanism of turfy soil affected by both nature factors and human activities, which is beneficial for the preservation of the turfy soil and also shows significant environmental ecological benefit.     

The accuracy of ground data used in remote-sensing investigations

A popular application of remotely sensed data is the estimation of an environmental variable at a point, with its extrapolation over a large area. For such work reliable ground data are required. This paper is divided into two parts; the first part is a review which looks at the considerations that are necessary for the collection of such ground data and finds that the spatial variability of the terrain has a profound effect on sample design. The second part illustrates these considerations by means of a case study in which airborne remotely sensed data are used to map the green-leaf-area index of grassland. It is concluded that even when large sample sizes were employed the error in ground data was greater than the error in remotely sensed data.     

Lithological mapping of Bela ophiolite with remote-sensing data

Bela ophiolite (BO), the largest ophiolite in Pakistan, is important to our understanding of the western margin of the Indian plate, particularly the collisional history of Indian and Eurasian plates. However, it is located in a remote location and has not been extensively studied. For example, no detailed geological map for this area exists. In this article, remote-sensing data were processed by different techniques that were selected based on reflectance spectroscopy data and compared with a local geological map for the upper unit of BO. False-colour images (Landsat Enhanced Thematic Mapper plus (ETM+) bands 7, 4, 2 in the red, green, blue (RGB)), colour-ratio composite images of Landsat ETM+ data (5/7, 5/1, 5/4 in the RGB), Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data (4/5, 6/7, 3/4 in the RGB) and mafic index images are used to discriminate different lithologies. Our results are consistent with published local geological maps. Based on these images and field data, we created a detailed geological map of BO. Two types of basalts are differentiated, and an ultramafic body around 66° 15′ E, 26° 39′ N is recognized for the first time in the lower unit of BO. Integrating all the available data, we suggest that BO formed in two episodes; the upper unit is a classic ophiolite sequence and is younger than the lower unit. First, it was obducted onto the lower unit, and then the two units obducted as one onto the Indian continental edge.     

Spatial improvement of human population distribution based on multi-sensor remote-sensing data: an input for exposure assessment

A spatial mismatch of hazard data and exposure data (e.g. population) exists in risk analysis. This article provides an integrated approach for a rapid and accurate estimation of population distribution on a per-pixel basis, through the combined use of medium and coarse spatial resolution remote-sensing data, namely the Defense Meteorological Satellite Program Operational Linescan System (DMSP/OLS) night-time imagery, enhanced vegetation index (EVI), and digital elevation model (DEM) data. The DMSP/OLS night-time light data have been widely used for the estimation of population distribution because of their free availability, global coverage, and high temporal resolution. However, given its low-radiometric resolution as well as the overglow effects, population distribution cannot be estimated accurately. In the present study, the DMSP/OLS data were combined with EVI and DEM data to develop an elevation-adjusted human settlement index (EAHSI) image. The model for population density estimation, developed based on the significant linear correlation between population and EAHSI, was implemented in Zhejiang Province in southeast China, and a spatialized population density map was generated at a resolution of 250 m?×?250 m. Compared with the results from raw human settlement index (59.69%) and single night-time lights (35.89%), the mean relative error of estimated population by EAHSI has been greatly reduced (17.74%), mainly due to the incorporation of elevation information. The accurate estimation of population density can be used as an input for exposure assessment in risk analysis on a regional scale and on a per-pixel basis.     

Retrieval of forest canopy attributes based on a geometric-optical model using airborne LiDAR and optical remote-sensing data

In the retrieval of forest canopy attributes using a geometric-optical model, the spectral scene reflectance of each component should be known as prior knowledge. Generally, these reflectances were acquired by a foregone survey using an analytical spectral device. This article purposed to retrieve the forest structure parameters using light detection and ranging (LiDAR) data, and used a linear spectrum decomposition model to determine the reflectances of the spectral scene components, which are regarded as prior knowledge in the retrieval of forest canopy cover and effective plant area index (PAI_e) using a simplified Li–Strahler geometric-optical model based on a Satellites Pour l'Observation de la Terre 5 (SPOT-5) high-resolution geometry (HRG) image. The airborne LiDAR data are first used to retrieve the forest structure parameters and then the proportion of the SPOT pixel not covered by crown or shadow K_g of each pixel in the sample was calculated, which was used to extract the reflectances of the spectral scene components by a linear spectrum decomposition model. Finally, the forest canopy cover and PAI_e are retrieved by the geometric-optical model. As the acquired time of SPOT-5 image and measured data has a discrepancy of about 2 months, the retrieved result of forest canopy cover needs a further validation. The relatively high value of R ² between the retrieval result of PAI_e and the measurements indicates the efficiency of our methods.     

Estimating chlorophyll a concentrations from remote-sensing reflectance in optically shallow waters

A multi-spectral classification and quantification technique is developed for estimating chlorophyll a concentrations, Chl, in shallow oceanic waters where light reflected by the bottom can contribute significantly to the above-water remote-sensing reflectance spectra, R_rs(λ). Classification criteria for determining bottom reflectance contributions for shipboard R_rs(λ) data from the west Florida shelf and Bahamian waters (1998-2001; n = 451) were established using the relationship between R_rs(412)/R_rs(670) and the spectral curvature about 555 nm, [R_rs(412) ? R_rs(670)]/R_rs(555)². Chlorophyll concentrations for data classified as “optically deep” and “optically shallow” were derived separately using best-fit cubic polynomial functions developed from the band-ratios R_rs(490)/R_rs(555) and R_rs(412)/R_rs(670), respectively. Concentrations for transitional data were calculated from weighted averages of the two derived values. The root-mean-square error (RMSE_log10) calculated for the entire data set using the new technique was 14% lower than the lowest error derived using the best individual band-ratio. The standard blue-to-green, band-ratio algorithm yields a 26% higher RMSE_log10 than that calculated using the new method. This study demonstrates the potential of quantifying chlorophyll a concentrations more accurately from multi-spectral satellite ocean color data in oceanic regions containing optically shallow waters.     

Development and application of a neural network based ocean colour algorithm in coastal waters

An algorithm for determining chlorophyll‐a concentrations in shallow, case II waters has been developed and applied to nearly six years of Sea‐viewing Wide Field‐of‐view Sensor (SeaWiFS) data in order to observe the general chlorophyll‐a patterns in a coastal estuarine environment. Due to the fact that the current empirical chlorophyll‐a algorithm (OC4) used to process SeaWiFS data breaks down in coastal waters, a neural network based algorithm was developed. The neural network in the study uses SeaWiFS remote sensing reflectance data paired with in situ chlorophyll‐a data in the Delaware Bay and its adjacent coastal zone (DBAC) from a number of different days and seasons in an effort to overcome the limitations of single day algorithms and simulated dataset algorithms. Although the neural network model (NN) in this study displayed some difficulty representing high chlorophyll‐a values, it showed significant improvement over the OC4 algorithm. The performance parameters of the NN were an r ² of 0.79, a root mean square (RMS) error of 3.69?mg m^?3 and a relative RMS error of 0.77. The NN was used to reprocess approximately six years of cloud free imagery of the DBAC from which the spatial and temporal variability of the chlorophyll‐a distributions in the DBAC were analysed. Time series of absolute chlorophyll‐a values for five stations along the central axis of the Delaware Bay were analysed using Fourier analysis techniques, from which chlorophyll‐a patterns were found to have a quasi‐annual period. Furthermore, the spatial distributions of the chlorophyll‐a patterns were analysed using a general climatology and monthly climatologies of normalized chlorophyll‐a values. The climatologies generally agreed with spatial distributions determined from historic ship‐based data. The study found that summer blooms in the mid‐estuary of the Delaware Bay may be more important than previously observed. This suggests that more frequent and synoptic measurements via satellite can reveal important new information about even well studied regions.     

基于卷积神经网络的遥感图像分类研究

遥感图像分类是模式识别技术在遥感领域的具体应用,针对遥感图像处理中的分类问题,提出了一种基于卷积神经网络(convolutional neural networks,CNN)的遥感图像分类方法,并针对单源特征无法提供有效信息的问题,设计了一种多源多特征融合的方法,将遥感图像的光谱特征、纹理特征、空间结构特征等按空间维度以向量或矩阵的形式进行有效融合,以此训练CNN模型。实验表明,多源多特征相融合能够加快模型收敛速度,有效提高遥感图像的分类精度;与其他分类方法相比,CNN能够取得更高的分类精度,获得更优的分类效果。     

Mapping bottom depth and albedo in coastal waters of the South China Sea islands and reefs using Landsat TM and ETM+ data

Optical models for the retrieval of shallow water bottom depth and albedo using multispectral data usually require in situ water depth data to tune the model parameters. In the South China Sea (SCS), however, such in situ data are often lacking or obsolete (perhaps from half a century ago) for most coastal waters around its islands and reefs. Here, we combine multispectral data collected by MODIS and Landsat to estimate bottom depth and albedo for four coral reef regions in the SCS, with results partially validated by some scarce in situ data. The waters in these remote regions are oligotrophic whose optical properties can be well derived from Moderate Resolution Imaging Spectroradiometer (MODIS) measurements when the waters are optically deep. The MODIS-derived optical properties are used to estimate the water column attenuation to the Landsat measurements over shallow waters, thus eliminating the requirement of model tuning using field measured water depths. The model is applied to four Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) images covering Pratas Atoll, Woody Island, Scarborough Shoal, and North Danger Reefs. The retrieved bathymetry around Pratas Atoll and North Danger Reefs are validated with some in situ data between 1 and 25 m. The relative difference and root mean square difference between the two measurements were 17% and 1.6 m, for Pratas Atoll and 11% and 1.1 m for North Danger Reefs, respectively. These results suggest that the approach developed here may be extended to other shallow, clear waters in the SCS.     

A five channel chlorophyll concentration algorithm applied to SeaWiFS data processed by SeaDAS in coastal waters

Chlorophyll-a concentration derived from the Sea-viewing Wide Fieldof-view Sensor (SeaWiFS) after applying the current SeaWiFS Data Analysis System (SeaDAS) processing tools appears to be higher than reality in coastal areas, particularly from late summer to early spring when optical properties of water are dominated by yellow substances and suspended matter. As a complement to the SeaWiFS standard procedure addressing clear water, empirical algorithms can bring immediate progress for observing the coastal domain. This paper proposes to modify the SeaWiFS Ocean Colour 4 band algorithm (OC4) by including the 412 and 555 channels. The effect of the suspended matter on the ratios used as inputs in OC4 is revealed by the 555 channel whereas the atmospheric over-correction and the absorption by yellow substances are related to the 412 channel. Based on a dataset located in the English Channel and on the continental shelf of the Bay of Biscay, a parametrization of the relationship between the OC4 ratio and the 412 and 555 bands has been empirically proposed for different chlorophyll concentrations. Application of a lookup table, relating triplets (OC4 band ratio, 412 and 555 bands) to chlorophyll-a concentration, provides realistic concentration maps likely to satisfy the needs of researchers involved in environmental surveying or fishery management.     

Towards a practical remote-sensing model of suspended sediment concentrations in turbid waters using MERIS measurements

A new empirical index, termed the normalized suspended sediment index (NSSI), is proposed to predict total suspended sediment (TSS) concentrations in inland turbid waters using Medium Resolution Imaging Spectrometer (MERIS) full-resolution (FR) 300 m data. The algorithm is based on the normalized difference between two MERIS spectral bands, 560 and 760 nm. NSSI shows its potential in application to our study region – Poyang Lake – the largest freshwater lake in China. An exponential function (R² = 0.90, p < 0.01) accurately explained the variance in the in situ data and showed better performance for the TSS range 10–524 mg l^?1. The algorithm was then validated with TSS estimates using an atmospheric-corrected MERIS FR image. The validation showed that the NSSI algorithm was a more robust TSS algorithm than the band-ratio algorithms. Findings of this research imply that NSSI can be successfully used on MERIS images to obtain TSS in Poyang Lake. This work provided a practical remote-sensing approach to estimate TSS in the optically and hydrologically complex Poyang Lake and the method can be easily extended to other similar waters.     

Monitoring of forest virtual water in Hunan Province,China, based on HJ-CCD remote-sensing images and pattern analysis

Virtual water has become an important part of global water supply and demand and has led to the globalization of water. Virtual water research most mainly focused on the field of agriculture. Minimal attention has been devoted to forest virtual water (FVW). To our knowledge, no research on the monitoring and analysis of FVW through remote-sensing technology has been conducted. In this study, based on object-oriented technology and through the use of 30 scenes from multi-temporal Chinese HJ-CCD images, we monitored FVW in Hunan Province, China, in 2010 and analysed the pattern of FVW. Results showed that the amount of FVW is large and greater than that of entity water. Hunan Province had 5.83 × 10¹¹m³ FVW in 2010, which was 3.09 times the amount of agriculture and livestock. FVW was thrice as large as the total entity water, 4.07 times the amount of surface water, and 14.57 times the amount of underground water in Hunan Province. The distribution of FVW in Hunan Province is uneven and presents a trend that gradually increases from northeast to southwest; nevertheless, the trend is reasonable and in favour of alleviating and optimizing the pattern of water resources. Our analysis indicated that we need to improve the cognition of virtual water and pay due attention to FVW from the perspective of water management and allocation. Our results also indicated that forest and woody products are water intensive. An efficient method of balancing FVW and other uses of water is thus required; control and management of water consumption in forests should also be implemented under the condition of protecting the environment. For China, woody forest products should be mainly imported to improve water-use efficiency and relieve the shortage of water. Meanwhile, remote-sensing technology is a useful tool, and Chinese HJ-CCD images are an important data source for the estimation of FVW.     

Optical closure for remote-sensing reflectance based on accurate radiative transfer approximations: the case of the Changjiang (Yangtze) River Estuary and its adjacent coastal area,China

Optical closure exercises are pivotal for evaluating the accuracy of water quality remote-sensing techniques. The agreement between radiometrically derived and inherent optical property (IOP)-derived above-water spectral remote-sensing reflectance R_rs(λ) is necessary for resolving IOPs, the diffuse attenuation coefficient, and biogeochemical parameters from space. We combined spectral radiometric and IOP measurements to perform an optical closure exercise for two optically contrasting Chinese waters – the Changjiang (Yangtze) River Estuary and its adjacent coastal area in the East China Sea. The final aim of our investigation was to compare two derivations of R_rs(λ): R_rs(λ), derived from radiometric measurements; and R_rs(λ), derived from simultaneous IOP measurements. Five subsequent steps have been taken to achieve this goal, including (1) estimation of the R_rs(λ) from radiometric measurements; (2) scattering correction for the non-water spectral absorption coefficient a_pd(λ); (3) estimation of the below-water spectral remote-sensing reflectance r_rs(λ) from IOPs measurements; (4) the estimation of the R_rs(λ) from the r_rs(λ) values; and (5) the comparison between the R_rs(λ) derived from radiometric and IOP measurements. All steps were realized by using both direct measurements and different models based on radiative transfer theory. Results demonstrated that the impact of the errors caused by the scattering correction procedure and conversion of radiometric quantities into R_rs(λ) may be rather significant, especially in the long-wavelength spectrum range. Nevertheless, spectral features were similar between these R_rs(λ) sets for all waters – from relatively clear to very turbid. Exploiting this fact allows use of the spectral reflectance ratios for remote sensing of the estuarine and coastal Chinese waters.     

A study of spatial and temporal dynamics of total ozone over Southwest China with multi-source remote-sensing data

This article describes a study of the spatial and temporal dynamics of total ozone over Southwest China using satellite-retrieved total ozone products from 1996 to 2008 and a ground-based Dobson spectrophotometer. The findings indicated that the value of total ozone (265.7 Dobson unit (DU)) over Southwest China is lower than the value (273.7 DU) over the adjacent region at the same latitude by about 8 DU, and is about 13.8 DU lower than the global average at the same latitude (279.5 DU), and that there is a distinctly low-value area due to the higher elevation. The relationship of total ozone and the elevation presents a negative correlation, the terrain being the main factor to affect this condition. In the long term, the variation of total ozone exhibits a slightly increasing trend from 1996 over this region. Total ozone presents an obvious seasonal change, with the largest value appearing in springtime and the smallest appearing in wintertime. The difference between the regional seasonal mean value of total ozone in springtime and wintertime is about 28 DU, although the difference between the maximum and minimum monthly total ozone throughout a year is up to 50 DU. There is a positive correlation between the variation of total ozone and relative humidity. Relative humidity may be an important factor impacting on the pattern of seasonal change of total ozone.     

Supervised vicarious calibration (SVC) of hyperspectral remote-sensing data

A full-chain process approach to extracting reflectance information from hyperspectral (HRS) data which is valid for all sensor qualities is proposed. This method is based on a mission-by-mission approach, followed by a unique vicarious calibration stage. As the HRS sensor's performance may vary in time and space, a vicarious calibration method to retrieve accurate at-sensor radiance values is necessary. In fact, vicarious calibration solutions usually rely on natural, well-known, bright and dark targets that are large in size and radiometrically homogeneous. Since such targets are not commonly found in the field for every mission and their spectral features can sometimes resemble artifacts in the corrected radiance, a new vicarious calibration approach is needed. This paper describes a new method that uses artificial agricultural black polyethylene nets of various densities as vicarious calibration targets that are set up along the airplane's trajectory (preferably near the airfield). The different densities of the nets combined with any bright background afford full coverage of the sensor's dynamic range. We show that these artificial targets can be used to assess data quality and correct at-sensor radiance within a short time. Several case studies are presented using Aisa-DUAL sensor data taken at different times from different locations. We found that even “lost data” (in terms of radiance drift) could be recovered by the suggested method. We term the suggested vicarious calibration approach supervised vicarious calibration (SVC) and demonstrate its performance in terms of spectral accuracy. The limitations of the method are also discussed but the overall conclusion is that the suggested procedure is functional, valuable and practical for sensors with questionable or uncertain laboratory-determined radiometric parameters.