遙測在環境監測之應用 Remote Sensing for Environmental Monitoring

Presentation on theme: "遙測在環境監測之應用 Remote Sensing for Environmental Monitoring"— Presentation transcript:

1 遙測在環境監測之應用 Remote Sensing for Environmental Monitoring
鄭 克 聲 台灣大學生物環境系統工程學系 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

2 Definition of Remote Sensing
The science and art of obtaining information about an object, area, or phenomenon through the analysis of data acquired by a device that is not in contact with the object, area, or phenomenon under investigation. 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

3 Energy Sources Modern remote sensing systems use electromagnetic energy as the source for image acquisition. 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

4 Primary Spectral Regions Used in Earth Remote Sensing
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

5 Energy Interactions In the atmosphere With earth surface features
Scattering Absorption With earth surface features Reflected Absorbed Transmitted 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

6 Paths of Energy Reaching the Sensor
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

7 Spectral Signatures Spectral reflectance curves
Spectral emittance curves (wavelength > 3.0 m ) 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

8 Satellite Remote Sensing Systems
Land surface observation Landsat SPOT ASTER IKONOS ALOS Quick bird Meteorological observation GMS NOAA series GOSAT Sea surface observation 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

9 Closer Look of Remote Sensing Images
GMS – Visible, IR SPOT – Panchromatic, MSS FORMOSAT-II IKONOS – Panchromatic GOSAT Airborne image 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

10 Remote Sensing Image Analysis
Pixel by pixel Signal processing in nature Require sophisticated softwares Employ mathematical/statistical methods Multitemporal and multispectral images are often used. Field data are essential. 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

11 Applications for Environmental Monitoring
土地利用與地表覆蓋判識 (Landuse/landcover interpretation) 地表變遷偵測與崩塌地辨識 (Land surface change detection and landslides identification) 作物生長狀況與水分逆境監測 (Monitoring crop growing condition and water stress) 區域蒸發散量推估 (Estimation of regional evapotranspiration) 水庫水質及優養狀態監測 (Monitoring of reservoir water quality and trophic status) 即時雨量預報 (Real-time rainfall forecasting) 瘧疾高風險區之劃定(Mapping areas with high risk of malaria) 乾旱監測與預警 (Drought monitoring and warning) 洪水平原與淹水區域劃定 (Mapping the floodplain and inundation zones) 森林火災監測 (Forest fire monitoring) 二氧化碳濃度監測 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

12 土地利用與地表覆蓋判識 (Landuse/landcover interpretation)
Multispectral images Spectral response patterns Can use both spectral and textural features for classification Supervised classification and unsupervised classification (cluster analysis) Illustrative example 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

13 Landcover (Taipei) 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

14 Landcover (Taipei) 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

15 Scattering of Training Pixels in 2-D Feature Space (IR and Green)
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

16 Landcover (Taipei) Partition of a 2-dimensional feature space (Infrared and Green; Indicator kriging) 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

17 Landcover (Taipei) Partition of a 2-dimensional feature space (Infrared and Green, Max. Likelihood) 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

18 Land Surface Change Detection and Landslides Identification
Multitemporal and multispectral Images Image-to-image registration Band ratioing (IR/R) of multitemporal images Band-ratio difference image Determine the percentage of LSC (histogram matching for band-ratio images) Determine the change detection threshold value (histogram of the band-ratio difference image) Usage of DEM data 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

19 IR band image pair after image-to-image registration
Image of 21/09/ Image of 01/10/1999 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

20 IR/R Band-Ratio Difference Image
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

21 Illustrative Grey-level Histogram of Band-Ratio Difference Image
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

22 Initial Landcover Change Identification
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

23 Identified Landslide Sites
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

24 Airphotos taken before and after Typhoon Toraji of an identified landslide site
Before Toraji After Toraji 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

25 Reservoir Trophic State Evaluation Using Landsat TM Images
Carlson Trophic State Indices (CTSI) Based on measurements of SDD, Chla and TP. Empirical relationships among the three inter-correlated parameters. CTSI is developed based on local and empirical relationships. Three indices corresponding to SDD, Chla and TP. 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

26 Carlson’s trophic state indices
Commonly used trophic states oligotrophic TSI  40; mesotrophic 40 < TSI  50; eutrophic < TSI. Direct application of CTSI in Taiwan is inadequate. 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

27 Te-Chi Reservoir trophic state indices
Reservoir TSI Estimation using Landsat TM images 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

28 Study Area 台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

29 TM-derived TSI(Chla) images
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

30 TM-derived TSI(TP) images
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

31 TM-derived TSI(SDD) images
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

32 Ttrophic state classes from upstream to outlet of the reservoir (10/01/1995)
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

33 Ttrophic state classes from upstream to outlet of the reservoir (22/07/1996)
台灣大學生物環境系統工程學系 Lab for Remote Sensing Hydrology and Spatial Modeling 遙測水文及空間模式研究室 Dept. of Bioenvironmental Systems Engineering, NTU

34 Path radiance calibration using in-situ measurements of radiometric control areas
An algorithm for path radiance calibration using surface reflectance measurements in two RCAs of the same topographic (unobstructed and horizontal) and landcover conditions was proposed.

35 RCA-based algorithm

36

37

38 Nonparametric classification algorithm
A Feature-Space Indicator Kriging Approach for Remote Sensing Image Classification (IEEE Transactions on Geoscience and Remote sensing) Nonparametric classification algorithm High classification accuracies (Producer’s, user’s, and overall accuracies) Account for spatial continuity in feature space, instead of geographic space.

39 Hypothesis-Test-Based Landcover Change Detection Using Multitemporal Satellite Images (Advances in Space Research) Two hypothesis-test-based change detection methods, namely the bivariate joint distribution method and the conditional distribution method, are proposed to tackle the uncertainties in change detection by making decisions based on the desired level of significance.

40 Image-differencing method
Multiples of standard deviation of DN difference Nelson (1983): k = 0.5~1 Ridd and Liu (1998): k = 0.9~1.4 Sohl (1999): k = 2

41 BVN and Conditional distribution

42 Detected Changes Area-A
α=1% α=5% α=10%

43

44 Area-B

45 Area-A

46 Area-B

47 Overall Accuracy vs Level of Significance
Area-A Area-B

48 A Multivariate Model for Coastal Water Quality Mapping Using Satellite Remote Sensing Images (Sensors) This study demonstrates the feasibility of coastal water quality mapping using satellite remote sensing images. Water quality sampling campaigns were conducted over a coastal area in northern Taiwan for measurements of three water quality variables including Secchi disk depth, turbidity, and total suspended solids. A spectral reflectance estimation scheme proposed in this study was applied to SPOT multispectral images for estimation of the sea surface reflectance. Two models, univariate and multivariate, for water quality estimation using the sea surface reflectance derived from SPOT images were established. The multivariate model takes into consideration the wavelength-dependent combined effect of individual seawater constituents on the sea surface reflectance and is superior over the univariate model.

49

50 Spatial distribution of secchi disk depth

51 Spatial distribution of turbidity

52 Spatial distribution of total suspended solids

53 Assessing the effect of landcover on air temperature using remote sensing images – A pilot study in northern Taiwan (Landscape and Urban Planning) NOAA AVHRR thermal images were used for surface temperature retrieval using the split window technique. SPOT multispectral images were used for landcover classification using the supervised maximum likelihood classification method. Through an inversion algorithm, landcover-specific surface temperatures were estimated. Locally calibrated relationships between surface and air temperatures with respect to different landcover types were developed using field data and used to yield average air temperatures over individual NOAA pixels.

54 Spatial variation of apparent surface temperature

55 Landcover conversion pattern

56 Under the prevalent landcover conversion, reducing the within-pixel coverage ratio of paddy fields from the maximum of 26% to none will result in an ambient air temperature rise of 1.7 to 3.1C. Forced landcover conversion contradict the existing landcover pattern and may cause complicated consequences. Addition resources allocation and incentives may need to be introduced in order to ensure a successful forced conversion.

57 Comparing landcover patterns in Tokyo, Kyoto, and Taipei using ALOS multispectral images (Landscape and Urban Planning) Understanding the landcover pattern of a region is essential for landuse planning and resources management. In this study ALOS multispectral images were used to compare landcover patterns in three study areas, namely Tokyo, Kyoto, and Taipei, of different degrees of urbanization. From the results of landcover classification, Shannon diversity index at cell level was used as a landscape metric for landcover pattern analysis. Existing landcover pattern of the three study areas were also compared by investigating cell distribution in a landcover coverage-ratio space. Both the landcover type richness and evenness are low in most of the Tokyo study area and built-up is the single dominant landcover type in almost all cells. In comparison, landcover patterns of the Kyoto and Taipei study areas are more diversified, with significant amount of cells having mixed and non-dominant landcover types. Kyoto is least urbanized and enjoys a good mixture of different landcover types. It was found that cell-average NDVI alone can be used for delineating areas with different dominant landcover types. Implementation of such method does not require an a priori LULC classification, and thus is particularly useful when good training data for LULC classification are not available.

58 Study areas and landcover images

59 Landcover vs SHDI images

60 Demonstration of higher SHDI in cells covering rivers and mountain foothills.

61 Landcover pattern in coverage-ratio space

62 Relationship between cell-average NDVI and cell-level SHDI.

63 Relationships between coverage ratios of different landcover types within individual cells of the Kyoto study area.

64 Areas of different dominant landcover types in the three study areas delineated using cell-average NDVI.

65 Cell-average NDVI alone can be used for delineating areas with different dominant landcover types. Implementation of such method does not require an a priori LULC classification, and thus is particularly useful when good training data for LULC classification are not available.