What is image space and feature space in remote sensing?

Answer 1

The image space is the 2D plane of the image where pixels are located. It represents the spatial space of the image. In other words, when we talk about the location of each pixel in an image, we are talking about image space. On the other hand, feature space is about the radiometric values assigned to each pixel. In case of a grey-scale imagery, only one radiometric value is assigned to each pixel. When we say an image is RGB or multispectral, then each pixel has several radiometric values that are stored in different channels (for instance there are 3 channels of Red, Green and Blue in an RGB image, so for a pixel we have 3 radiometric values). Feature Space is the space of these radiometric values; the radiometric values of each pixel can be plotted in that space and you can create a feature space image. Last example, an RGB image has a 3 dimensional feature space while it still has a 2D image space.

Answer 2

Image space refers to the pixel values present in a remote sensing image, reflecting the intensity of electromagnetic radiation recorded by the sensor. Feature space, on the other hand, involves extracting meaningful information from the image data by using a set of features or attributes to represent different objects or phenomena in the scene. Feature extraction techniques help to characterize and classify the image data for various applications in remote sensing.