I used landsat 8, with band combination 2,5,4 to show the cultivated areas, and the urban areas, also the roads, and any settlements
Hi,
I am not familiar with the region, and your question does not clarify if you'll be working with visual interpretation of the images, or a more formal statistical image classification procedure. Anyway, using only bands from the visible spectral domain is a waste of the information available form Landsat-8. If your aim is to create RGB color composite images for visual interpretation, I recommend you use one band from the visible domain (possibly RED, which is less affected by atmospheric effects than the blue and green bands, loaded in the BLUE color plane), one from the NEAR-INFRARED domain (loaded in the GREEN color plane), and another one from the SHORTWAVE INFRARED (loaded in the RED color plane). Although the colors will not be intuitive, your ability to distinguish between different land surface types (including water) will be greatly enhanced.
I agree with Jose. Landsat 8 images are wonderfully loaded with spectral information, so visually interpreting them is wasting the information. Otherwise, I would use band 1 as well, since you mentioned about coastal region.
What classification system are you using? Is there a reason for only using bands 2,3,4 as opposed to additionally adding the near and short wave infrared bands? Would creating a data fusion/resolution merge image with the pan band help your efforts?
This combination will work as Reza Rafee mention. In coastal region it is possible there are many small water bodies( Wetlands) and salt pans. .if band 5 induces in your study it gives a more distinguishable land cover map.
Dear Pushparaj, I've uploaded a screenshot which compares a Landsat 8 scene of a coastal region around 250 miles south of the state of Karnataka, India acquired in the month of February this year. I presume your area might have some similarity. The image to the left is made up of band combinations 654 according to the description provided by Professor Pereira and the image to the right is a simple Natural Color Image with band combinations 432. You can see how visually appealing the image to the left is where you can quite clearly distinguish between water bodies, vegetation, soil, built up area etc. with the shades of blue, green and purple colours. This image can be a good source for your classification. But as Dr. Gomez, Scott and Reza have inquired, you need to mention quite specifically what you are going to classify because there are many possible combinations of bands depending on the purpose of your study. If you are interested in using 'Coastal Band' (Band 1) then I have a whole section of Q&A in my RG page. It's always good to try out different band combinations and check which features are enhanced.
Credits & Sources: ERDAS Imagine 9.2, MS Paint.
http://blogs.esri.com/esri/arcgis/2013/07/24/band-combinations-for-landsat-8/
http://earthobservatory.nasa.gov/Features/FalseColor/page6.php
http://landsat.usgs.gov/L8_band_combos.php
http://earthobservatory.nasa.gov/IOTD/view.php?id=84218&src=fb
I think you can use more bands to make classification,it will let you have a better result.You can also use band-5,band-6 in LANDSAT-8 data.
Hello, it depends also on which land cover you want to focus on. Distinction between vegetatation is made well by composite 6,5,4 (5,4,3 ETM+), I mostly use this combination of bands.
Also it is helpful for another purposes of classification and this composition is used in many studies.
Wish good luck!
h.
Yes, Well. I use landsat-8 image. I use image fusion techniques. And apply to Pan Band covers range of 0.50-0.68. and the same range covered by Band B-2,G-3,R-4. And my study area covers portion of coastal(sea), river, and buildings etc, by using the above mentioned band can i go for general classification , (0r) what kind of feature can be classified by using the above band.
- Badges
- Science topic
- Similar topics
- Correction