1

I was learning eCoginition by tutorial, but still, not familiar with some advanced funcionality, such as stretching raster data. the distribution of one band in my image

As you can see, the values are condensed in a small portion of the range. So,the goal is to stretch it to the scale of 0-255 using eCog. And the stretching process should make actual change to the data, not just for visualization. Any ideas? Thanks in advance.

1 Answer 1

3

The eCognition software provides some basic data visualization functions to facilitate visual interpretation of your input image. You may perform raster stretching operations like histogram, standard deviation, linear or manual stretching of individual or all loaded bands. These operations are meant for visual appreciation and NO actual changes are made to the input bands.

Once you load an image in the eCognition, it is assumed that basic data processing operations such as radiometric, geometric calibrations are already performed and data is ready for analysis. The eCognition is certainly NOT right place to pre-process your data. You may opt for available image processing software either commercial like ArcGIS, ERDAS, ENVI etc. or Open source solutions like QGIS, SAGA, GRASS GIS etc. to work with.

We might not required pre-processing of our data when our objective is limited to perform image segmentation and subsequent landcover classification on an independent image. The raw DNs would be sufficient for the intended purpose. However, in many cases, we are supposed to take into account data calibration to make proper use of data. E.g. for change detection analysis within eCognition, we must have to ingest geometrically and radiometricaly corrected pair of images to produce reliable outcome.

1
  • Sincerely thank you for detailed answer! It's clear that the stretching is unneccesary in my case. Commented Mar 3, 2020 at 5:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.