INVESTIGATION OF INTER-FRAME CORRELATION OF WEBCAM-GENERATED CHAOS

Abstract

The essence of research of this work is sequences of random numbers (RNS) obtained from consecutive frames of a webcam, which can be practically used to create a hardware RND generator.

The problem under consideration was to calculate the similarity level of RNS s obtained from consecutive webcam frames and to estimate the looping period of RNS generation.

The obtained results confirmed the assumption of a high level of chaos and randomness in the generation of RNS by an ordinary webcam. In a set of serial capture frames (delay - 40 milliseconds), the photodiode matrix registers changes that the human eye cannot detect. In conditions of absolute darkness, changes in brightness values were fixed in almost 60% of the pixels of the matrix. Which is 10% higher than the level of the requirement for the cryptoresistance in terms of the avalanche effect parameter. This fact can be explained by the stochastic nature of the interaction of photons with the sensor material and thermal noise. Thus, we can talk about a clear positive correlation of the level of chaos in frames with the growth of the delay time between frames.

The specialty of the research is that for the stringency of the experiment, frame generation was carried out in complete darkness (illumination - 10^-4 lux) and a uniformly illuminated (illumination of 200 lux) white surface. Tests of the ordinary webcamera under extremly boundary conditions give a complete picture of the determination of chaos in the generation of RNS.

The result of this research makes it possible to design an algorithm that will form the basis of the development of a hardware laboratory RNS generator without the involvement of exotic equipment.