I have to use to OTSU algo on T-cap Greenness component in order to find out the threshold above/below which i can figure it out as vegetation/agriculture/urban greenery.
As i have used MATLAB...(with "graythresh" function) and tried ENVI Image Change Workflow... (where i get a change detection difference map with threshold technique of OTSU).
What is the expected out for this...?
Hi Atiqa,
When you apply a threshold to an image, you basically are segmenting the image into regions. Thresholding would produce a binary image based on the grey tones, classifying pixels falling below the threshold to zero and the rest above the threshold to 1. You can even adjust the threshold to somehow minimize the overlap between the region.
I hope this helps.
Thats the issue Eric. I tried thresholding and clipped to my study area in ENVI, but the values ranges from -200 to 178 not from -1 to 1, as ideally expected.
Sometimes if two peaks are not clearly identified in histogram then it may lead to error. I had used a r package (EBImage) for otsu on a raster. You may try that too.
Yeah... two un-clearly identified peaks, always effect the value... As OTSU generates a single values against the histogram.
you can select the threshold of the two images and you can even add the third and set a condition which will be suitable for both. Example " you can have algorithm like this: Image1 or Band 1 > threshold Image1 orB1 & Image2orB2 < threshold B2 & B3 < threshold B3
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping