A television camera for forensics with an image fusion function.
Vyacheslav Mikhailovich SMELKOV, PhD in Engineering, Associate Professor
An article was published in the Special Equipment magazine in issue No. 4 for this year about a technical solution for a television camera designed to monitor forensic objects with increased accuracy by increasing the image resolution at the rate of low-frame decomposition.
The latter assumes that the vertical and horizontal scan rates and the element-by-element frequency readings are significantly slower and, accordingly, lower compared to the broadcast standard according to GOST 7845-92.
In this case, the TV camera should ensure, first of all, the formation of maximally undistorted video information for its subsequent input into the computer.
Adhering to the thesis that most objects of forensic diagnostics are objects with a high density of individual components, but are in a static state, the author of this work proposes to expand the functionality of a low-frame TV camera by forming a combined image signal at its output.
When applied to two spatially separated objects, the new function is provided by forming two undistorted video signals from a single CCD matrix in each active low-frame line.
The image combination function is often encountered in television systems operating in the American or European broadcasting standard.
However, it should be noted that the formation of a video signal of a combined image is necessarily accompanied by an exchange of resolution for the sizes of the images “joined” within the visible raster.
For example, according to an English patent [1], a combined image generated at the output of an analog video system maintains maximum horizontal resolution, but there is also a partial loss of video information from each of the television cameras, since only the central sections of the two component images are transmitted in the combined image.
On the other hand, a modern digital quadroscope “packs” two full (lossless) images into the horizontal raster size, but the clarity of the combined image, compared to the original, is two times worse.
Below is a technical solution for a low-frame television camera equipped with a function for combining two images, but without the noted shortcomings.
The TV camera is based on a CCD matrix with a “frame transfer” organization with two symmetrically located horizontal registers [2, p. 112].
It should be noted that such a circuit organization of the photodetector was specially proposed by domestic developers in the early eighties for the operation of a CCD camera in a broadcast or compatible standard.
This organization allows to select either the accumulation section or the memory section as a photo target and, depending on the belonging of the admissible dark current defects (centers of increased thermal generation-recombination) to one or another section, to significantly reduce the “contribution of this interference to the output video signal of the matrix.
Thanks to this possibility, a geometrically large “column” defect can always be exchanged for a small “point” one!
A distinctive feature of the proposed TV camera is the low-frame rate of self-scanning scanning of the CCD matrix, the pulse exposure mode of the photodetector and the use of its accumulation section and memory section as two photo targets simultaneously.
The video signal of a TV camera line contains a number of counts (pixels) equal to the total number of elements in the registers of the CCD matrix, and the video signal of a frame contains a standard number of lines and a 4/3 format, which coincides with the format of an individual photo target.
The structural diagram of a TV camera, the device of which is recognized as an invention [3], is shown in Fig. 1.
It contains an optical unit (1); a CCD matrix (2); four level converters (3), (4), (5) and (6); a control signal generator (7); two video amplifiers (8) and (9); a video signal processing unit (10).
Fig. 1. Structural diagram of a television camera according to the invention [ 3]
The optical unit (1) has two main tasks:
- implementation of the pulsed illumination mode of two controlled objects (“left” and “right”) using two synchronously operating pulsed light sources;
- changing the scale of each of the optical images in the horizontal direction.
Optical scaling, which acts as an optical “stretching” horizontally for input images, is performed by anamorphic lenses.
Note that the industrial implementation of such lenses with the necessary horizontal anamorphic coefficients currently does not cause technical difficulties [4].
The guiding optical elements (1-5), (1-6) and (1-8) perform a parallel translation of each of the two optical axes.
The reversing optical element (1-7) rotates the image on the target (2-2) relative to the image on the target (2-1) by 180 degrees.
It should be added that, to simplify the drawing of the structural diagram of the television camera, the course of the frame rays is not shown in Fig. 1.
In industrial production of the optical unit (1), it is advisable to make the inverting optical element (1-7) in the form of a Dove prism or a Pechan prism.
Let us consider the operation of a low-frame television camera. The images of the “left” and “right” objects in the pulse exposure mode are projected through the optical unit (1) onto the targets (2-1) and (2-2) of the CCD matrix.
The pulse exposure mode, provided by synchronously operating pulsed light sources (1-1) and (1-2), is carried out so that during the reverse stroke of the frame scan, both optical channels are open, and during the forward stroke, they are closed.
Control of blocks (1-1) and (1-2) is performed using a frame blanking pulse supplied from the output of the control signal generator (7).
The anamorphization coefficients of lenses (1-3) and (1-4) over the frame field are equal to 0.5.
Therefore, a pulsed image of the “left” object, stretched twice horizontally, is sent to section (2-1), and a pulsed image of the “right” object, stretched twice in the same direction, is sent to section (2-2).
By the end of the frame blanking interval on photo targets (2-1) and (2-2), the formation of charge reliefs for the “left” and “right” control objects, respectively, is completed.
During the forward stroke of the frame scan, charge packets are read out one by one with the optical channels closed (in the dark) in section (2-1) to the output register (2-3) and from section (2-2) to the register (2-4).
During the first half of the active part of each line, all horizontal pixels of section (2-1) are read, and during the second half, all pixels of section (2-2) are read.
It is important to note that the doubled frequency of charge packet transfer in the registers, due to the low-frame decomposition mode, remains significantly lower than the maximum sampling frequency of the CCD array, for which it is necessary to take into account image signal distortions due to transfer inefficiency.
As a result, the video signals at the first and second outputs of the CCD matrix contain video information with a restored (undistorted) scale for the “left” and “right” objects of control.
Then, through the buffer video amplifiers (8) and (9), the video signals from the CCD matrix are fed to the unit (10), where the procedure of their preliminary processing and combining (“stitching”) is performed.
At the output of the TV camera, a combined video signal is formed at a low frame rate, in each active line of which the image signals from the “left” and “right” objects of control are sequentially transmitted in full.
It should be noted that the use of one CCD matrix for two video channels ensures the identity of operating modes, the identity of the change in characteristics during aging, and the stability of the mutual geometric characteristics of the two image targets.
If M is the number of pixels for the first or second registers of the CCD matrix, then the horizontal resolution of the combined image (N x) in television lines is maintained along the line equal to:
N x =M 1/k, where k is the frame format (4/3).
As already mentioned, all other things being equal, this result is twice as great as the capabilities of a modern digital square.
Literature
- UK Patent No. 1438293. MKI2 H04N 7/08, NKI H4FD2A, D30T. Television system for transmitting two or more images over one cable. Published 3.06.76.
- Kuznetsov Yu.A., Shilin V.A. Microcircuitry of LSI on charge-coupled devices. — M. “Radio and Communications”, 1988.
- Patent of the Russian Federation No. 2065257. MKI6 H04N 5/225. Low-frame device for forming an image signal.//V.M. Smelkov./B.I. – 1996 — No. 22.
- Cards of anamorphic attachments and blocks for shooting widescreen films./Review of domestic and foreign film lenses for shooting conventional, widescreen, wide-format and 16-mm films. Compiled under the editorship of F.S. Novik. Moscow, 1969, p. 165.