Huang Jingfeng

GIS-based logistic regression method for landslide susceptibility mapping in regional scale

Landslide susceptibility map is one of the study fields portraying the spatial distribution of future slope failure susceptibility. This paper deals with past methods for producing landslide susceptibility map and divides these methods into 3 types. The logistic linear regression approach is further elaborated on by crosstabs methods, which is used to analyze the relationship between the categorical or binary response variable and one or more continuous or categorical or binary explanatory variables derived from samples. It is an objective assignment of coefficients serving as weights of various factors under considerations while expert opinions make great difference in heuristic approaches. Different from deterministic approach, it is very applicable to regional scale. In this study, double logistic regression is applied in the study area. The entire study area is first analyzed. The logistic regression equation showed that elevation, proximity to road, river and residential area are main factors triggering landslide occurrence in this area. The prediction accuracy of the first landslide susceptibility map was showed to be 80%. Along the road and residential area, almost all areas are in high landslide susceptibility zone. Some non-landslide areas are incorrectly divided into high and medium landslide susceptibility zone. In order to improve the status, a second logistic regression was done in high landslide susceptibility zone using landslide cells and non-landslide sample cells in this area. In the second logistic regression analysis, only engineering and geological conditions are important in these areas and are entered in the new logistic regression equation indicating that only areas with unstable engineering and geological conditions are prone to landslide during large scale engineering activity. Taking these two logistic regression results into account yields a new landslide susceptibility map. Double logistic regression analysis improved the non-landslide prediction accuracy. During calculation of parameters for logistic regression, landslide density is used to transform nominal variable to numeric variable and this avoids the creation of an excessively high number of dummy variables.

Rice yield estimation using remote sensing and simulation model

Remote sensing techniques have the potential to provide information on agricultural crops quantitatively, instantaneously and above all nondestructively over large areas. Crop simulation models describe the relationship between physiological processes in plants and environmental growing conditions. The integration between remote sensing data and crop growth simulation model is an important trend for yield estimation and prediction, since remote sensing can provide information on the actual status of the agricultural crop. In this study, a new model (Rice-SRS) was developed based mainly on ORYZA1 model and modified to accept remote sensing data as input from different sources. The model can accept three kinds of NDVI data: NOAA AVHRR(LAC)-NDVI, NOAA AVHRR(GAC)-NDVI and radiometric measurements-NDVI. The integration between NOAA AVHRR (LAC) data and simulation model as applied to Rice-SRS resulted in accurate estimates for rice yield in the Shaoxing area, reduced the estimating error to 1.027%, 0.794% and (−0.787%) for early, single, and late season respectively. Utilizing NDVI data derived from NOAA AVHRR (GAC) as input in Rice-SRS can yield good estimation for rice yield with the average error (−7.43%). Testing the new model for radiometric measurements showed that the average estimation error for 10 varieties under early rice conditions was less than 1%.

In situ hyperspectral data analysis for pigment content estimation of rice leaves.

Analyses of the correlation between hyperspectral reflectance and pigment content including chlorophyll-a, chlorophyll-b and carotenoid of leaves in different sites of rice were reported in this paper. The hyperspectral reflectance of late rice during the whole growing season was measured using a Spectroradiometer with spectral range of 350–1050 nm and resolution of 3 nm. The chlorophyll-a, chlorophyll-b and carotenoid contents in rice leaves in rice fields to which different levels of nitrogen were applied were measured. The chlorophyll-a content of upper leaves was well correlated with the spectral variables. However, the correlation between both chlorophyll-b and caroteniod and the spectral variables was far from that of chlorophyll-a. The potential of hyperspectral reflectance measurement for estimating chlorophyll-a of upper leaves was evaluated using univariate correlation and multivariate regression analysis methods with different types of predictors. This study showed that the most suitable estimated model of chlorophyll-a of upper leaves was obtained by using some hyperspectral variables such asSDr,SDb and their integration.

Monitoring the Sequential Cropping Index of Arable Land in Zhejiang Province of China Using MODIS-NDVI

The sequential cropping index of arable land is important agricultural information. The aim of this article is to monitor and analyze the parameter, and offer reference for agricultural production. The cropping index of arable land in Zhejiang Province, China from 2001 to 2004 was calculated using the second order difference based MODIS (moderate resolution imagine spectroradimeter) vegetation data from NASA (National Aeronautic and Space Administration) in America and the land use map with a scale of 1:25 000. It was found that the peak of the time series of the NDVI curve indicated that the ground biomass of crops reached the maximum, and fluctuated with the crops growing processes such as sowing, seeding, heading, ripeness, and harvesting within one year. Thus, the sequential cropping index was defined as the number of peaks of the time series of the NDVI curve. The sequential cropping index of all cities in Zhejiang Province, China was worked out. It is seen from the spatial distribution that the cropping index in the southwest Zhejiang Province is larger than that in the northeast. As for the temporal distribution, the sequential cropping index decreased from 2001 to 2003, whereas it increased slightly from 2003 to 2004. However, the index of arable land was relatively low, as far as the geographic position and climatic resource were concerned, and the potential of the sequential cropping index was great.

Estimation of vegetation biophysical parameters by remote sensing using radial basis function neural network

Hyperspectral reflectance (350∼2500 nm) data were recorded at two different sites of rice in two experiment fields including two cultivars, and three levels of nitrogen (N) application. Twenty-five Vegetation Indices (VIs) were used to predict the rice agronomic parameters including Leaf Area Index (LAI, m2 green leaf/m2 soil) and Green Leaf Chlorophyll Density (GLCD, mg chlorophyll/m2 soil) by the traditional regression models and Radial Basis Function Neural Network (RBF). RBF emerged as a variant of Artificial Neural Networks (ANNs) in the late 1980’s. A large variety of training algorithms has been tested for training RBF networks. In this study, Original RBF (ORBF), Gradient Descent RBF (GDRBF), and Generalized Regression Neural Network (GRNN) were employed. Results showed that green waveband Normalized Difference Vegetation Index (NDVIgreen) and TCARI/OSAVI have the best prediction power for LAI by exponent model and ORBF respectively, and that TCARI/OSAVI has the best prediction power for GLCD by exponent model and GDRBF. The best performances of RBF are compared with the traditional models, showing that the relationship between VIs and agronomic variables are further improved when RBF is used. Compared with the best traditional models, ORBF using TCARI/OSAVI improves the prediction power for LAI by lowering the Root Mean Square Error (RMSE) for 0.1119, and GDRBF using TCARI/OSAVI improves the prediction power for GLCD by lowering the RMSE for 26.7853. It is concluded that RBF provides a useful exploratory and predictive tool when applied to the sensitive VIs.

Integration of remote sensing data and simulation model to estimate rice yield

Remote sensing techniques have the potential to provide information on agricultural crops quantitatively, instantaneously and above all nondestructively over large areas. Crop simulation models describe the relationship between physiological processes in plants and environmental growing conditions. The integration between remote sensing data and crop growth simulation model is an important trend for yield estimation and prediction. In this study, a new model (Rice-SRS) was developed based mainly on the ORYZA1 model and modified to accept remote sensing data as input from different sources. The model can accept three kinds of NDVI data: NOAA AVHRR (LAC)-NDVI, NOAA AVHRR (GAC)-NDVI and radiometric measurements NDVI. The integration between NOAA AVHRR (LAC) data and simulation model as it applied by Rice-SRS resulted in accurate estimation for rice yield in the Shaoxing area, characterized by reducing the estimating error to 1.027%, 0.794% and (-0.787%) for early, single and late season respectively. Utilizing NDVI data derived from NOAA AVHRR (GAC) as input in Rice-SRS can offer good estimation for rice yield with average error (-7.43%). Testing the new model for radiometric measurements was successful, since the average estimation error for 10 varieties under early rice conditions was less than 1%.

The natural grassland production monitoring and estimation in the north of Xinjiang using remote sensing

In this paper, the data of spectral observations and the grass yield of natural grassland in the north of Xinjiang from 1991 to 1994 are used for analyzing the spectral vegetation index characteristics, studying the relation between the ground spectrum and satellite remote sensing data and establishing the grass yield monitoring model and the remote sensing monitoring and estimating model of natural grassland to the north of Xinjiang.