Radiometric Errors is the distortion in pixel brightness values.
The distortions suffered by images in recorded values at different pixel locations are referred to as radiometric distortions.
Broadly classified into the two categories listed below
A. Internal errors
B. External Error
A. Internal errors
The devices themselves cause internal errors. Because of their systematic nature, these types of errors are also known as systematic errors.
Based on laboratory calibration or in-flight measurements, these errors can be modelled, identified, and corrected.
B. External Error
External errors are caused by natural phenomena that vary in space and time, and are thus referred to as non-systematic errors.
External factors such as atmospheric disturbances and steep terrain undulations can cause radiometric and geometric errors in remote sensor data.
Correction of radiometric errors necessitates knowledge of EMR principles as well as the interactions that occur during the data acquisition process.
The radiometric correction can benefit from terrain information such as slope and aspect, as well as advanced information such as the scene’s bi-directional reflectance distribution function (BRDF). Radiometric correction procedures can be time consuming and problematic at times.
Internal Errors Corrections
A. Random Bad Pixels :
When a pixel’s received signal is not recorded by an individual detector.
As a result, bad pixels may appear at random.
Shot noise occurs when a scene contains a large number of random bad pixels.
The image appears to have many dark poke marks due to the short noise. In general, these bad pixels have values in one or more bands ranging from 0 to 255 (in 8-bit data).
Shot noise is removed by identifying pixels in a given band that are either 0 (black) or 255 (white) in the midst of radically different pixel values.
The average pixel values of their respective eight neighboring pixels are then used to replace the noise pixels.
B. Line or Column Dropouts
If an individual detector in an electro-mechanical scanning system ( Landsat 7 ETM+) fails to function properly, a blank row with no details of features on the ground may be seen.
The bad line or column is commonly referred to as a line or column drop-out because it has brightness values equal to zero or some constant value that is independent of terrain changes.
how to correct ?
In general, this is an irreversible loss of information because there is no way to recover data that was never obtained.
However, it is possible to improve data visual interpretability by introducing estimated brightness values for each bad scan row (or column) by replacing it with the average of rows (or columns) above and below (or to the left and right).
This concept works because adjacent pixels frequently have similar pixel values.
C. Line Start Problems
Is that the scanner does not begin recording as soon as a new row begins.
It is also possible that the sensors will place pixel data at inconvenient locations along the scan line.
All of the pixels in a scan line could be systematically shifted to the right by one pixel.
This is known as a line-start problem.
If the line start problem is always associated with a fixed number of column horizontal bias, it can be corrected with a simple horizontal adjustment.
External Error Corrections
The three steps for correcting external errors are as follows:
A. Conversion of DNs to at sensor spectral radiance
B. The conversion of spectral radiance to apparent reflectance
C. The removal of atmospheric effects
