The processing images of infrared band are produced in order to identify zones of interest or objects. The choice of algorithms and processing tools depends on the task of both reductions in visibility or object detection. These tasks are opposite in matter, but they have a similar algorithmic solution, comprises in need of highlighting the areas of luminosity, while the luminosity is considered as a characteristic of the contrast of the object from the background, relatively in the spectrum, which raises the meaning of problem to determine the boundaries of the object. The object detection task in the infrared range of the spectrum is different to a similar problem in the visible spectrum on several grounds, such as alter-native distribution of contrasts, the presence of self-luminescence of the observed object, which allows you to define the objects only through their differences in emissivity of the surface in the absence of incident radiation and temperature gradients. Thus, the infrared images include a variety of data missing in the images obtained in the visible range. So that two main features are followed by: data saturation of the image; and pro-visual interpretation task in due to the need for special methods and approaches to analyze their appropriacy. Image processing is the task of processing the brightness of the matrix obtained by thermal imager. For the purpose of convenient image mathematical processing it is converted into a grayscale image which corresponds to a linear change in brightness temperature change that allows using more simplified mathematical algorithms to reduce the process of handling interval and the required calculating resources. Mostly, the transferring process is a standard function for images in the visible spectrum, which promotes the emergence of the problem, which consists in the use of different color palettes that do not have a clear definition of the order of color gradient of thermal imagers scales from different manufacturers of apparatus, as well as the possibility of defining «by eye» which palette is used. Therefore, high-quality treatment and recovering the data of interest in the image is made only in heat visor software products. In the case of the thermal image with the unknown one, or, in the absence of specialized software products the processing will be performed with an accuracy on the order of color gradient. The solution to this problem lies in the infrared image itself, the color of the applied field, and specifications of maximum values. The proposed algorithm produces a pixel by pixel analysis of the color palette inside the set borders of the color fields and creates a matrix comparing the used palette with palette of White Hot given finite values shown in the picture. Then it steps the pro-conversion of the entire image to the palette of White Hot to have the correct temperature markers that makes it possible to determine the exact value of the temperature at any pixel image. As a result, the temperature data set is storing exactly in new format, and further processing is carried out on the converted image.
Flury B. Principal points. Biometrika. N.77. 1990. P. 33-41.
Pytev JU.P. Zadachi morfologicheskogo analiza izobrazhenijj // V sb. Matematicheskie
metody issledovanija prirodnykh resursov Zemli iz kosmosa / Pod red. V.G. Zolotukhina. M.: Nauka. 1984.
Meer P., Mintz D.,
Rozenfield A., Kim D.Y. Robust regression methods
for computer vision: a review // Int. J. Comput.Vision. 1991. V. 6(1). P. 59-70.37.
Cunikov A.JU. Obrabotka izobrazhenijj v infrakrasnom diapazone spektra (Monografija).
JAroslavl: Remder. 2011. 60 s.