Experimental Dependence of X-ray Film Noise on the Average Level of Its Transparency

A. S. Moskaleva

(Presented by the scientific seminar of the Department of Industrial and Medical Electronics)

The sensitivity of photographic materials, including X-ray films, is limited by noise, by which in photography is meant the nonuniformity of blackening of the emulsion layer. The nonuniformity of blackening of a photographic material, detected at low magnification and called, in this case, granularity, is very similar to random fluctuations of current or voltage occurring in an electrical circuit [1, 2]. The difference is that in the theory of electrical circuits the fluctuations of electrical quantities are considered as functions of time, whereas in the theory of photographic images the fluctuations of blackening are functions of space. However, the spatial and temporal dependences of a one-dimensional random process, such as the nonuniformity of blackening in photography, owing to its isotropy [3], are related to one another through the Fourier transform [4]. As in television, the evaluation of the magnitude of noise in photography is conditioned by the need to determine the number of distinguishable gradations of blackening. Noise in photography blurs the sharp outlines of images, which leads to an increase in the minimally perceptible difference in transparencies on the film.

In order to determine the possibility of filtering such noise, experiments were carried out with samples of uniformly exposed and developed X-ray films of the RT-1 type. Samples of various transparencies were scanned with a slit aperture of \(10 \times 50\ \mu^2\); the light signal passing through the sample was converted by a photomultiplier into an electrical signal, which was recorded by a digital voltmeter in discrete form. From the obtained realization the correlation function was calculated, and then the spectral power density of the process. As a result, the dependence of the spectral power density of the film noise at various transparencies was obtained. The figure gives curves of the dependence of the spectral power density \(G(\omega)\) on the average transparency of the samples \(T\). As a result of analysis of the obtained results, the following conclusions can be drawn:

1. The character of the curves obtained by us for X-ray films coincides with the character of such curves obtained for photographic films, the data for which were published in [5].

2. The main power of the noise spectrum is concentrated at low frequencies; in this connection, the problem of studying—

…of the possibility of filtering the low-frequency components of fluctuations in the transparency of the photographic material.

1. The power of one and the same frequency component of the process changes when passing from one transparency to another.

In the figure: vertical axis \(G(\omega)\); horizontal axis \(T\); \(A = 10 \times 50 \mu^2\). Curves are labeled \(\omega = 0\), \(\omega = 0.1\), \(\omega = 0.2\), \(\omega = 0.4\), \(\omega = 0.6\), \(\omega = 0.8\), \(\omega = 1.0\).

Fig. 1. Dependence of the noise spectrum of RT-1 X-ray film on the mean transparency over the sample

REFERENCES

1. P. Fellgett. J. Opt. Soc. Amer. 43, 271, 1953.
2. E. L. O’Neill. J. Opt. Soc. Amer. 48, 945, 1958.
3. K. V. Vendyurovskii, M. A. Aizman, I. G. Minkevich. Uspekhi nauchnoi fotografii, vol. XI, 1966.
4. A. A. Kharkevich. Spectra and Analysis of GITTL, Moscow, 1962.
5. F. J. B. Wall and B. G. Steel. J. Phot. Sceinse 12, pp. 34—46, 1964.