Capabilities

Methods of thematic processing of satellite images allow you to study images in detail and receive vector layers with attribute information, including in automatic mode. In The IMAGE MEDIA CENTER software package implements a wide range of tools for thematic processing of satellite images. 

Image analysis in false colors includes:

• the use of various combinations of color channels;
• method of greatest similarity;
• use of color spaces (RGB, CMYK, Lab, HLS, HSB).

The image shows the original image from Landsat-8 to the territory Republic of Sakha. In the second window – image obtained in pseudo colors (7-6-4).

The brightness value of each pixel of the index image is formed by performing mathematical operations in which values are used as parameters the brightness of each pixel from different channels of the image.
Depending on the purpose of the study different indexes are used:

• vegetation indices;
• soil indices;
• water indices;
• snow indexes;
• custom indexes.

For the formation of index images in the IMC PC, the tool "Calculator channels", which allows the user to compose any mathematical formulas with using image channels.

To calculate the surface temperature a number of auxiliary calculations are performed (spectral radiation intensity, surface brightness temperature, spectral emissivity), conversion of values ​​to degrees Celsius and a universal temperature scale is used. To detect thermal anomalies, calculation of surface temperature, selection of areas of anomalous temperatures, vectorization objects and filling with attributive information.

Before the start of clustering, it is not known how many and what objects are in the image, and after clustering, it is necessary to decipher the resulting classes in order to determine whether what objects they correspond to. Thus, classification without training is applied:

• if it is not known in advance which objects are in the picture;
• the image contains a large number of objects (more than 30) with complex boundaries;
• can also be used as a preliminary step before supervised classification.

For supervised classification, reference areas are used, which defines an operator based on their belonging to a certain class of objects. For the next recognition as training samples, the values ​​of the pixels of the reference areas in different spectral ranges. Thus, each pixel of the image belongs to to a certain class based on sequential comparison with all created standards. At controlled classification, information classes are first determined, and then, based on them, spectral.

The image shows the result of the classification of the underlying surfaces with image training of "Resurs-P" spacecraft, "KShMSA-SR" equipment.

For supervised classification, reference areas are used, which defines an operator based on their belonging to a certain class of objects. For the next recognition as training samples, the values ​​of the pixels of the reference areas in different spectral ranges. Thus, each pixel of the image belongs to to a certain class based on sequential comparison with all created standards. At controlled classification, information classes are first determined, and then, based on them, spectral.

The image shows the result of the classification of the underlying surface with the training of the image of the Resurs-P spacecraft, the KShMSA-SR equipment.

The main quantities to be changed in the spectral analysis are wavelength, intensity of the reflected signal and the spatial coordinate of the studied surfaces. The IMC software implements the following spectral analysis methods:

• correlation with/without regard to amplitude;
• binary encoding;
• spectral-angular mapping;
• orthogonal projection of a subspace.

The figure shows a hyperspectral image from the Resurs-P spacecraft, spatial spectrogram built by a row of pixels, as well as spectral graphs, received from various types of objects.