RU2538340C1 - Method of superimposing images obtained using different-range photosensors - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: in the method, each odd-numbered line corresponds to an infrared image and each even-numbered line corresponds to a television image with further interlinear exchange of parts of opposite pixels based on an algorithm, in which each pixel is divided in proportion into three parts, after which two parts of the i-th pixel of the k-th line is summed with the remaining portions of the i-th pixels of the (k-1)-th and (k+1)-th lines.

EFFECT: high information value of the resultant image.

4 dwg

Description

The present invention relates to electronic equipment, namely to two-dimensional image filtering.

The known method "Integration of digital halftone television and thermal imaging images" (RF patent No. 2451338, IPC G06T 5/00, 12/23/10).

The essence of the method of combining digital halftone television and thermal imaging images is to obtain the original images, combining images based on criteria summation for each pixel, forming the resulting image and normalizing the brightness range of the image. The original digital television and thermal imaging images are obtained, for example, by digitizing the analog signals from the television and thermal imaging sensors using analog-to-digital converters, the images are complexed based on a certain summation for each pixel, i.e. generate the resulting image in the form of an array of brightness values of the complex image of a given size, normalize the brightness range of the image.

In the above method, the pixels of the resulting image contain information determined either by the IR range or the visible frequency range, which does not give a significant increase in information content.

Also known is the “Method for processing infrared images” (RF patent No. 2437153, IPC classes G06T 5/50, H04N 5/33, 07/09/07), which consists in expanding the functionality of processing an infrared image due to user-controlled processing during image capture.

A method for processing an infrared image includes the steps of processing an infrared image to provide a background portion of an infrared image and a detailed portion of an infrared image; scaling the background portion and / or the detail portion to provide a level of the detail portion relative to the level of the background portion; combining the background part and the detailed part after scaling to provide a processed infrared image; and save the processed infrared image.

The disadvantage of this method is that the infrared image has lower detail than a television.

Closest to the claimed method is the "Method of combining images obtained using multi-range photosensors" (RF patent No. 2435221, IPC class G06T 3/00, 10.12.07).

The essence of the method is as follows: the luminous flux reflected from the studied objects is formed into separate images and recorded in a common image. The general image is used to judge the relative position of objects. Preliminarily, the luminous flux is divided into two fluxes, an image of the visible spectrum is formed from the first flux, an infrared image is formed from the second flux, the pixels of the first and second images are read, and the overall image of these pixels is alternately recorded in the memory. On odd lines, pixels of the first and second images are recorded sequentially, and on even lines, pixels of the second and first images are sequentially recorded. EFFECT: increased accuracy of determining the mutual arrangement of image fragments and the ability to calculate their position parameters.

The disadvantage of the above method is that each pixel contains information of either a television or infrared image, therefore, the information content of the total image is increased by no more than two times.

The main task, which is aimed by the claimed method of combining images is to increase the information content of the image.

The technical result is achieved due to the fact that in the developed method, the light flux reflected from the objects being studied is formed into separate images and recorded in a common image, according to which the relative position of the objects is judged, moreover, the light flux is previously divided into two streams, from the first stream they are formed image of the visible spectrum, from the second - the image of the infrared spectrum, read the pixels of the first and second images and alternately record the total image of these pixels in memory, and each odd line corresponds to an IR image, and each even line corresponds to a TV image (Fig. 1) with further inter-line exchange of parts of opposite pixels according to an algorithm in which each pixel is divided into three parts, after which two parts of the i-th pixel of the k-line are summed with half fractions of the ith pixels of the k-1st and k + 1st rows, (FIG. 2). The final pixel position is shown in FIG. 2.

The effect of combining with preservation of all fragments of both the first and second images is obtained due to the fact that the image of the visible spectrum and the image of the infrared spectrum are obtained from the same stream, that is, these images are taken from one angle and the lines are written sequentially in the combined image first and second images.

Ergonomic efficiency of the developed method is achieved by alternately recording strings of TV and IR images in a common image with further division of pixels into three parts and interline exchange of these parts for pixels of the same number in rows.

Figure 1 shows the process of alternately recording strings of TV and IR images. Figure 2 is an enlarged block diagram of pixels with interpolation shown. Figure 3 - the final result of the process.

n _k is the amount of information contained in the IR image, n _t is the amount of information contained in the TV image, 2n _kt is the amount of information contained in the resulting image, which contains information about both TV and IR images.

As a result of the formation, the information content after combining two images in one doubled compared to the original images, which is confirmed by the formula (1):

$${\text{n}}_{\text{k}}+{\text{n}}_{\text{t}}={\text{2n}}_{\text{kt}}\text{. (one)}$$

Then there is a process of line-to-line interpolation, as a result of which each pixel of the resulting image will represent half of its pixel and a quarter of neighboring ones.

Thus, the pixel will have information not only about itself, but also about two neighboring pixels, that is, the information content of the pixel will increase by three times.

n ^* + n ^< + n ^> = 3n,

where n ^* is the amount of information of an unchanged half of the pixel, n ^< is the amount of information received by a quarter of the pixel on the left, n ^> is the amount of information received by a quarter of the pixel on the right, 3n is the final amount of pixel information.

The final formula for informativeness (2):

$${\text{2n}}_{\text{kt}}\text{* 3n}=\text{6n}\text{. (2)}$$

We carry out a comparative analysis of the first and second formulas, we obtain:

. $$\frac{\left(nk+nt\right)*3n}{nk+nt}=\frac{6n}{2n}=3$$

.

As a result, the developed method increases the information content of the prototype three times (since it increases the information content of the final image relative to the original two times), and relative to the original image - six. Thus, ergonomic efficiency is achieved because each pixel of the generated image contains information received from different-range sensors.

An example of a combined image is figv, which combined the TV image (Fig.4A) and the IR image (Fig.4b).

The results of the search for known technical solutions in this and related fields of technology in order to identify features that match the distinctive features of the claimed invention from the prototype showed that they do not follow explicitly from the prior art.

The inventive method is useful and workable and can be used in medicine, and more specifically in thermography, for the diagnosis of skin diseases, as well as in search work.

Claims (1)

A method of combining images obtained using various photosensors, in which the light flux reflected from the objects under study is formed into separate images and recorded in a common image, according to which the relative position of the objects is judged, the light flux being previously divided into two streams from the first stream form the image of the visible spectrum, from the second - the image of the infrared spectrum, read the pixels of the first and second images and alternately record the total image of these pixels in memory characterized by the fact that each odd line corresponds to an infrared image, and each even line corresponds to a television image with further inter-line exchange of parts of opposite pixels according to an algorithm in which each pixel is divided into three parts in proportion, after which two parts of the ith pixel k- the lines are summed with the remaining fractions of the ith pixels of the k-1st and k + 1-st lines.

Images