Discussions & Conclusions

In ArcGIS,

· Need to carefully follow the steps & procedures

· Take extra caution on the pixel and threshold values

· Stretched colours are useful of familiarising the data sets

· Take plenty of iterations to correctly pick out the flood extents

· Small pixel values usually denote flood water extents

· Use the base maps wisely

· GIF animation helps to visualise the changes of the flood extents Intensity and coherence analyses also show changes over the period

· Need to take plenty of iterations to correctly pick out the flood extents and generate the finest flood maps

Experiment discussions

Comparisons

In comparisons, intensity images can accurately show the flood extends but not in real colours. True colours shows with real colours but confused with other features. Infrared images shows the floodwater extent correctly as same as true colour images which do not show flood underneath vegetation. SES shapefile is fast in response to emergency events, but the flood boundaries are rough and exclude the non-flooded areas in between.

From the above result comparisons, the infrared image (top right) and intensity image (bottom left)  both show the flood extents correctly while true colour image (top left) confuses users with features apart from floodwater. SES flood extent (bottom right) shows clear boundaries as well but the lines are just a rough record with helicopters flying above them.

Intensity images can also show the density of the flood. Denser blue colour denotes large flooded area like farmlands and less dense extents shows the flood over towns and vegetation.

On the other hand, intensity analysis shows positive and negative changes of the area of interests. But it does not show the flood boundaries because the it can be no change in those areas.

Coherence analysis is the opposite which shows the flood boundaries but did not tell any positive or negative changes.

Moreover, SAR analysis can be compared with optical imagery as well. In general, SAR allow diversify analyses in computer while optical imagery is good in visualisation as true colour images show the flooded areas in real colours and infrared images shows the flood extent with surrounded features at the same time.

In  conclusion, different methods and techniques are applied according to different intentions and purposes. Each sensor contains different benefits and limitations, depends on the requirements of the users.

In general, ground-based technique is relatively slower, highly concerned in safety issues but good in accuracy. Airborne techniques allow fast results and access to remote areas. But it requires high operating costs on aerial platforms and are subject to weather condition. Space borne techniques can operate in any weather condition. It’s economical, fast and allow uses to monitor the flood continuously. But there’s a chance to miss the flood peak, depending on the revisiting time of the satellites.

So at the end, using space borne techniques especially radar is recommended for flood mapping, in terms of site access, safety issues, size of the flood events, labour usages, analyses of the flood extents, costs and issues of data acquisitions.

Concluding remarks

University of New South Wales

Integrated Remote Sensing For Flood Mapping

Type

Pros

Cons

Intensity images

Accurately shows the flood extents

No blockages: cloud/vegetation canopies

Pure colour based, not in real colour

True colour images

Show the extent in real colours

Confused with other features

Do not show flood underneath vegetation

Infrared images

Show the floodwater extent correctly

Not in real colours

Do not show flood underneath vegetation

SES shapefile

Fast mapping method

Flood boundaries are rough

Exclude the non-flooded areas in between

Limited to aerial platforms flying ranges