What are the differences and connections between geometric correction, orthorectification and geometric registration?

The differences and connections between them are:

Geometric correction is divided into different levels, and orthographic correction can be said to be the highest level of geometric correction. Compared with geometric correction, geometric correction mainly aims at the error of the data itself, adds real corresponding geometric coordinate information to the data itself, and restores the authenticity of the data itself.

Generally speaking, geometric correction is to eliminate the geometric distortion caused by atmospheric transmission, the sensor itself, the curvature of the earth and other factors, mainly to correct or give the image plane coordinates. Orthographic correction not only has the function of conventional geometric correction, but also can correct the distortion of the image caused by topographic relief according to DEM, which will add elevation information to the image.

Compared with geometric correction, geometric correction mainly aims at the error of the data itself, adds real corresponding geometric coordinate information to the data itself, and restores the authenticity of the data itself. Geometric registration is the geometric relationship between graphics, that is, an image is registered with a reference image.

So geometric correction is more like a data processing in the early stage, and geometric registration is more like a data processing in the later stage. The principle of image registration is essentially the same as orthorectification, which is to geometrically match images with different phases, different bands or different types, so that the images have unified geographical coordinates and pixel spatial resolution.

There are two kinds of geometric correction:?

1. Geometric coarse correction: correction of distortion causes.

2. Accurate geometric correction: using control points to make geometric correction, using a mathematical model to approximately describe the geometric distortion process of remote sensing images, using some corresponding points between distorted remote sensing images and standard maps (that is, control point data pairs) to obtain this geometric distortion model, and then using this model to make geometric distortion correction, regardless of the cause of direct distortion.

System geometric correction data refers to the data processed by radiation correction and system-level geometric correction, that is, PCD ephemeris parameters are extracted from satellite downlink data and then used for system-level geometric correction of TM data. For geometric correction processing, its geographical positioning accuracy will be greatly improved. The geometric correction of the system adopts fast correction method, and the image-based distortion is decomposed into nominal distortion caused by scanning and satellite and distortion caused by seven kinds of perturbation, which are satellite offset respectively, with good geometric accuracy.

Extended data:

The key to the geometric correction of the system is to establish the conversion relationship between the line of sight and the uncorrected image plane to the corrected image plane in the fixed coordinate system of the earth. The calibration process is divided into the following four steps:

(1) establish LOS in the earth's fixed coordinate system.

(2)LOS is projected to the plane of geodetic coordinate map;

(3) establishing the mutual conversion relationship between the input plane and the output plane;

(4) Finding the corresponding relation of the output plane points in the input plane;

(5) resampling is performed in the whole output plane.

References:

Baidu encyclopedia-geometric correction