top of page

Image Transformation

Single-band transformation is based on algorithms independently applied to each one of the bands acquired by a sensor. The correction procedure of radiometric errors is the main single-band transformation, where the cause is corrected for each band. In geometric corrections and image registration, each pixel of the transformed image is assigned the radiometric value corresponding to the nearest pixel in the original image, or, more commonly, to the average of a variable number of pixels next to that one. Other single-band transformations are the operations carried out in digital image processing techniques, such as the modification of the histograms enhancing the contrast in image visualization, and the digital operators, or filters. The operators, operating through mobile windows, assign the radiometric value of each pixel of the filtered image in relation to the original values and those of the neighbors.

In multispectral transformations, bands can be combined to produce, from the originals, synthetic bands, also called pseudo-bands. These transformations can generate new images with higher information used in the interpretation phase or as processing before an automatic classification

Classification of the image transformation

Single band Multispectral

Radiometric correction Algebraic operation with bands

Geometric correction Vegetation indices (VI)

Image to image registration Tasseled Cap

Histogram modification Principal components analysis (PCA)

Digital filter Classification

Source: Basics of Geomatics

For more Visit:

7 views0 comments
bottom of page