Contribution form remote sensing
in updating bathymetric chart
Data Processing
Depth Points and Pixel Values for Calibration
The water depth measurements were corrected for tidal height variations and the readings were reduced to the level of the Lowest Astronomical Tide (Chart Datum). The tidal data relevant to the analysis of tidal correction were taken from the published tide tables (Royal Malaysian Navy, 1999). The number of points chosen for depth calibration and validation were 175 and 43 respectively.
The image pixel positions were related to the geographical coordinates through rectification to the bathymetric chart using selected ground control points and second-degree polynomial transformation equations. The locations of these sounding points were then transformed into image scan-line and column-numbers by using the corresponding calibrated transformation equations obtained from the rectification analysis. The pixel intensities (digital values) at these points were then obtained for bands 1, 2 and 3. For each pixel location, a 3 by 3 sample window was used for the extraction of the pixel value. The actual water height at any point during image acquisition was determined by adding the height of the tide above chart datum to the corrected depth obtained earlier. The data quality for bathymetric application using each data set was validated by observing the distribution of the data on the plot of pixel intensity versus depth (Figure 1).
Determination of Deep-water Radiance
One of the prerequisites of the algorithms listed above is the value of the deep-water radiance for each band. The procedure involved finding the pixel value of a deep-water area within the scene. The deep-water radiance is normally (but not always) characterized by the lowest digital number in each band.
Due to the high degree of uncertainty in obtaining these values from the image inspection approach, we determined this parameter by optimizing equation (2) through regression analysis. The iteration process determined the optimum value of L
si. This method was adopted in this study because the calibration accuracy was found to be better than the regression results using the deep-water radiance obtained by image inspection.
The results of the analysis gave the L
si values in terms of DNs for the first 3 bands at about 80, 25 and 8 for the TM data and the SPOT scene gave values of 60, 20, and 10 respectively.
Calibration Regression
A least squares minimization was performed to calculate the best values of the coefficients in the algorithm. The results using all the data points showed high RMS error due to the presence of many outliers that degraded the calibration accuracy. To minimize the root-mean-square (RMS) error, the points having high residuals were discarded. After a sufficient number of iterations to discard these outliers only 168 points were retained. The RMS deviation error and correlation coefficient, R, for each data set were noted.
Inspection of the regression results indicates that the accuracy increases from the single-band to the three-band algorithms. The proposed algorithm produced some improvement over the conventional three-band method. This algorithm was used for further analyses. The results using the proposed algorithm are displayed in Table 1.
Validation of the Calibrated Algorithm
The validation data sets, which were not used in the calibration regression analysis, were then used for depth evaluation. Each calibrated algorithm was applied to the respective remotely sensed data set. Scatter plots of depths from chart versus calculated depths were made and the correlation coefficients determined (Table 2).
Generation of Water Depth Maps
The water depth readings from echo sounding were used to plot the suggested updated bathymetric map of the area (Figure 2). Assessment for the contribution of remote sensing data for bathymetric charting was made using the present data sets. Each calibrated algorithm was applied to the corresponding multispectral image on a pixel-by-pixel basis. The land and cloud pixels were masked out using the infrared bands. The density?sliced depth images were geometrically corrected to the same coordinate system of the generated bathymetric map. The generated maps were colour-coded for display of the different depth contours (Figure 2).
Results and Discussion
The newly bathymetric map shows close similarity in depth patterns with the generated depth maps using remote sensing data. Observation of the plots of intensity versus depths and the results of calibration and validation for all the data sets reveal that TM and SPOT have comparable performances. TM and SPOT displayed nearly similar depth patterns. The principal error arose from the effect of water turbidity. Non-uniformity in the water turbidity and bottom reflectance caused additional complexities with this technique.
Conclusion
New depth information was obtained from the techniques employed in the present study. The results seggest that the existing bathymetric maps need to be updated. Remote sensing data can provide a valuable contribution in bathymetry. However, the environmental conditions may have a major influence in the accuracy of the technique.
Acknowledgment
This study is partially supported by the National Space Development Agency of Japan through the joint NASDA-ESCAP project entitled 'National Capacity building for sustainable environment and natural resources management through research and studies on the use of ADEOS data' and the Malaysian Government IRPA Grant no: 08-02-056011. Thanks are extended to Universiti Sains Malaysia.
References
- Benny, A. H. and Dawson, G. J., 1983. Satellite imagery as an aid to bathymetric charting in the Red Sea. Cathographic Journal, 20: 5-16.
- Lee, Z. P., Carder, K. L., Mobley, C. D., Steward, R. G., and Patch, J. S., 1999. Hyperspectral remote sensing for shallow waters: 2 . Deriving bottom depths and water properties by optimization. Applied Optics, 38: 3831-3843.
- Lyzenga, D. R., 1978. Passive remote sensing techniques for mapping water depth and bottom features. Applied Optics, 17: 379?383
- Peredes, J. M. and Spero, R. E., 1983. Water depth mapping from passive remote sensing under a generalized ratio assumption. Applied Optics, 22: 1134-1135.
- Philpot, W. D., 1989. Bathymetric mapping with passive multispectral imagery. Applied Optics, 28: 1569-1579
- Royal Malaysian Navy, 1999. Tide Tables, Vol. I, 2000. Hydrographic Branch, Royal Malaysian Navy.
- Stove, G. C., 1985. Use of high-resolution satellite imaging in optical and infrared wavebands as an aid to hydrographic charting and coastal engineering. Proceedings of the Technical Programme, Electronics in Oil and Gas, London, pp. 509-530.
