study area is large. Additionally, it is hard to control all sets of training samples with
the same quality. Therefore, effective image normalization methods should be
developed to let training sample selection process much easier. During classification process parameters determination of using SVM is critical for obtaining the
best result. Also, it is difficult to determine which combination of the parameter
setting is the most superior.
4.7 Recommendations
Based on both the superiorities and limitations of this study, some potential future
studies are proposed here. With the global coverage of Landsat data, the time-serial
change detection method can also be applied to other cities or metropolitan areas or
even global scale to detect the LULC dynamic change. As for urban area analysis
by using remote sensing data, an effort can be given into improving the urban area
classification result. Since machine learning classifiers can deal with high dimensional dataset, various input features can be integrated together to investigate their
effectiveness of improving the classification result. Moreover, the time-serial
remote sensing data, GIS data and socio-economic data can be also incorporated
to generated more accurate urban growth model. Furthermore, the accessibility of
long-term Landsat record also makes it possible to detect time-serial dynamic
change of different land cover types, such as dynamic change of forest cover,
glacier, watershed and coastline.
4.8 Summary
This chapter gives an introduction of long-term change detection from a different
perspective. By using long-term high-dense Landsat dataset, specific detailed LULC
change dynamics can be extracted based on per-image classification. Compared to
coarsely multi-temporal change detection, long-term trajectory of LULC dynamic
change can provide higher-order complexities of LULC change. Information, such as
acceleration and deceleration can be analyzed. The detailed long-term change processes are very valuable information for planners and governments to understand the
causes and consequences of LULC change to make more appropriate and effective
regulations and policies for better planning and environmental management.
References
Abd El-Kawy OR, Rød JK, Ismail HA, Suliman AS (2011) Land use and land cover change
detection in the western Nile delta of Egypt using remote sensing data. Appl Geogr
31(2):483–494
4 Long-Term Change Dynamics Using Landsat Archive for the Region of Waterloo. . .
83
the same quality. Therefore, effective image normalization methods should be
developed to let training sample selection process much easier. During classification process parameters determination of using SVM is critical for obtaining the
best result. Also, it is difficult to determine which combination of the parameter
setting is the most superior.
4.7 Recommendations
Based on both the superiorities and limitations of this study, some potential future
studies are proposed here. With the global coverage of Landsat data, the time-serial
change detection method can also be applied to other cities or metropolitan areas or
even global scale to detect the LULC dynamic change. As for urban area analysis
by using remote sensing data, an effort can be given into improving the urban area
classification result. Since machine learning classifiers can deal with high dimensional dataset, various input features can be integrated together to investigate their
effectiveness of improving the classification result. Moreover, the time-serial
remote sensing data, GIS data and socio-economic data can be also incorporated
to generated more accurate urban growth model. Furthermore, the accessibility of
long-term Landsat record also makes it possible to detect time-serial dynamic
change of different land cover types, such as dynamic change of forest cover,
glacier, watershed and coastline.
4.8 Summary
This chapter gives an introduction of long-term change detection from a different
perspective. By using long-term high-dense Landsat dataset, specific detailed LULC
change dynamics can be extracted based on per-image classification. Compared to
coarsely multi-temporal change detection, long-term trajectory of LULC dynamic
change can provide higher-order complexities of LULC change. Information, such as
acceleration and deceleration can be analyzed. The detailed long-term change processes are very valuable information for planners and governments to understand the
causes and consequences of LULC change to make more appropriate and effective
regulations and policies for better planning and environmental management.
References
Abd El-Kawy OR, Rød JK, Ismail HA, Suliman AS (2011) Land use and land cover change
detection in the western Nile delta of Egypt using remote sensing data. Appl Geogr
31(2):483–494
4 Long-Term Change Dynamics Using Landsat Archive for the Region of Waterloo. . .
83