The main approaches to the development of an automated optoelectronic system of ground-space monitoring, providing secure processing of visual information under conditions of information competition, are considered. The achievements have theoretical and practical value when solving the problem of information performance provision in different areas of interests demanding visual information processing.
Global informatization and information competition stipulate for introduction of a rational methodology of building efficient automated optoelectronic system of ground-space monitoring (AOES GSM) having high values of accuracy, promptness, reliability, stability, survivability, as well as a high degree of information security to ensure functioning in an aggressive information environment. One of the central problems in the field of automated image processing is ensuring high precision of solving specific tasks resistant to various detrimental aggressive factors, with a clearly defined and studied range of application.
through the use of a problem-oriented version of the integrated
ICSD-approach
(“informational-cybernetic-synergeticdidactic”) [
a conceptual-logical model of AOES GSM (Fig. 1) [
This structure is presented as a complex of functional subsystems: measurement (P1), observation (P2), identification (P3), management decision-making (P4), centralized coordination (P5), information exchange (P6) and information protection (P7) required for functioning under the conditions of information competition and ensuring the necessary protection in the course of information processing.
actions at time t: functional R(t), external target X(t) and external coordinating X'(t), for which corresponding responses Q(t), Y(t),
localization and classification of objects of interest in photo and
video data in relation to various background-target situations.
Difficulties in solving these problems arise due to the following:
information losses when projecting a 3D scene (exposition) onto
the image plane; presence of noise in the image; changes in the
scene exposition; complex shapes of objects; changes in the
object shape; partial or complete overlaps and obturations of
scene objects; complex path of object motion; object goes
beyond the frame boundaries or appears in the frame; relative
motion of the camera; real-time processing requirements etc.
[
III.
from the set M of possible methods according to algorithm A.
А: ∈ .
measurement subsystem (P1) and the coordination subsystem (P5).
problems under conditions of informational competition. An effective algorithm A for choosing a method of action is formed in the result of learning in actual conditions and the narrowing of the set of uncertainty H.
Н → 0, А = { , }.
V{σR, σA, σsh}
I(x,y)
AOES GSM R{rs B, n, δ, x, φ, Hr, Нef, β}
Stabilization of the visual flow
Q={QDLK, QIE, Qа}
P{d, k, ρ}
D{ω} Learning of neural
networks Q={Qc, QIE, Qа}
Prompt processing of visual information Q={Qc, QDLK, QIE} Evaluation of information efficiency and functional diagnostics Q={Qc, QDLK, QIE, Qа} (Hн-H(Q*))=max)
Fig. 2 Reorganization scheme and mapping of the multilevel AOES GSM on the conceptual-logical model
elements at the input to the set Y of output results for a ∈ given method of action selected from the set M under conditions of uncertainty H.
G: X × M × H → Y.
determines the rule for displaying the set Y of output results for a ∈ given method of action selected from the set M into the set of quantities R associated with the characteristics of the system operation quality
К: M × Y → .
observation subsystem (P2) and the identification subsystem (P3).
on the received analytical information. The main subsystem functioning at this level is the decision making subsystem (P4).
At the same time, the required degree of security and safety
of information arrays is supported by the information protection
subsystem (P7) continuously at all levels. Information security
includes ensuring the reliability (noise immunity and noise proof
features), confidentiality (secrecy, accessibility and imitation
resistance), safeguarding (integrity, availability) of privileged
visual information. Thus, information security consists in the
ability to prevent accidental or deliberate distortion or
destruction, disclosure or modification of information arrays in
the information base [
COMPETITION
by a combination of properties that determine the degree of its
compliance with the goals of processing. In the general case, one
can distinguish internal quality indicators that characterize the
system content richness and external quality indicators that
determine the security of the system [
Internal quality indicators are responsible for such properties
of information processing as materiality - a property of
information that allows you to maintain value over time and
includes indicators of completeness, ensuring rationality of
management, and relevance, responsible for compliance with
the dynamic state of objects, as well as cumulativity - a property
that allows reflecting reality in a small informational array and
includes selectivity indicators providing the qualification choice
of a limited number of information units from a large-volume
informational array, and homoformism, which allow
aggregating large informational arrays into a small number of
informational units [
External quality indicators provide confidentiality properties for the reliability and safeguarding of information. The confidentiality property determines the status of the information external availability and is characterized by accessibility indicators (determining the degree of differentiation in the use of information arrays), secrecy (consisting in the ability to withstand the disclosure of the information meaning), imitation resistance (determining the degree of information protection from implementation and consisting in the ability to prevent disinformation from entering the system).
The reliability property characterizes the degree of correspondence of real information units to their true value and is determined by the parameters of noise immunity and noise proof features. The safeguarding feature is the possibility of targeted use of visual information arrays and the ability to prevent violation of the integrity of information arrays during the system operation.
Target characteristic: Safety
Кef max Accuracy КA
Promptness
Кpr Reliability
Speed Timeliness
Validity of decisionmaking
Quality factor КQ Functioning under fault conditions
Stability Кstab
Survivability
Кsurv
Stability Managyeabilit Adaptability Continuity
Performance indicator Кef = Σλif(КA, Кpr, КQ, Кstab, Кsurv) max
System general technical requirements for information algorithmic support Promptness of methods
Multi-level solutions
Invariance to a class of solved problems
Ef ective use of technical figures
Statistical indicators
Compatibility
Assessment of AOES GSM efficiency Functional diagnostics Reliability Methods
Implementations
Hardware
Statistical indicators
Time indicators
Ensuring the required efficiency: Кef max
In the general case, the technological process of processing the visual information flow is a series of sequential actions aimed at analyzing information and bringing it to the desired form at the system output. All the steps for processing visual information, as a rule, are a sequential removal of noninformative components from the image and the selection of informative ones to solve the tasks. The following three main technological operations can be distinguished in the technological process of processing the input information flow. The first one is reception, including the processes of formation and registration of images, as well as their preliminary compression and restoration. The second one is the interpretation resulting in intellectual processing, generalization of the received data, control and decision making.The third one is the communication operation responsible for receiving and transmitting information flow. The main operations, in turn, include various subprocesses that allow to form, register, compress and restore images, summarize information obtained as a result of intellectual processing and make management decisions based on the same, as well as transmit and save information.
In order to ensure security, partially processed visual information containing certain data necessary to ensure further restoration of full significance is transmitted to the communication operation instead of the general array of visual data. For example, in case of using neural network methods of decoding visual information, learning samples are generated, and the weights of the neural network obtained in the process of learning are transmitted for further work through the information channel instead of arrays of learning images.
of AOES GSM are indicators of information accuracy KA, promptness Kpr, Q-factor KQ, stability Kstab, survivability of Ksurv, each of them corresponding to a certain importance indicators 0 ≤ ≤ 1 ∑ = 1, , i = 1, ..., 5. The main qualities of these characteristics are the maximum use of the potential capabilities of AOES GSM. Information security of AOES GSM is defined as a function of private performance indicators Кef=Σ i f(КA, Кpr, КQ, Кstab, Кsurv) max.
The considered proposals for ensuring secure processing of information in AOES GSM are of theoretical and applied importance in solving the problems of developing effective automated optoelectronic systems of special (intersectoral, etc.) designation, providing processing of multi-aspect visual information. They can significantly increase the efficiency of using existing methods and tools for detection, localization and classification of images and, as a result, improve the quality of visual information recognition, ensure the confidentiality of information processing, and also increase the safeguarding and reliability of situation assessment under conditions of intense information competition and continuously changing environment.