I resize the Multispectral image to perform IHS-to-RGB transformation. I used the Nearest Neigbour Interpolation method when resizing. While researching this issue, I saw that this is the best option to prevent spectral contamination, but I do not know exactly what the definition of spectral contamination means. I searched, but could not find detailed information on this topic.
What is the meaning of spectral contamination in image processing?
"Spectral contamination" is used for designating a few phenomena connected with spectral analysis in optics, acoustics, analytical chemistry. I have never heard of these words to be used in connection with resizing of digital color images.
The gray-color images can acquire colored pixels in the course of resampling, one user of Adobe Photoshop calls this effect a "color variations".
In computer graphics, the colored edges appear in screen font glyphs rendered with LCD and OLED displays as the result of subpixel rendering. This technique is also used to enhance the resolution of all image types on layouts which are specifically designed to be compatible with subpixel rendering.
You can find online articles where research is done to detect the spectral contamination with the help of multispectral image cameras, but this is the spectal contamination connected with spectral analysis in optics, acoustics, analytical chemistry and it is not related to the color variations resulting from resampling the digital images.
