Аннотация
This paper addresses the problem of secure encryption of images and binary data through the integration of chaotic dynamic systems and modern symmetric cryptographic primitives. The growing volume of multimedia data in telecommunication and information systems necessitates the development of encryption methods that ensure high resistance to statistical and differential cryptanalysis while accounting for the intrinsic structural properties of visual data. A hybrid encryption algorithm is proposed, combining chaotic permutation and diffusion mechanisms with authenticated symmetric encryption based on the AES-256 algorithm in GCM mode. Chaotic sequences are generated using a logistic map, whose parameters are deterministically derived from key material via a key derivation function (Scrypt), enabling an expanded key space and high sensitivity to key variations. The algorithm implements a multi-stage transformation architecture, including chaotic preprocessing and subsequent authenticated encryption, ensuring both confidentiality and data integrity. The performance and cryptographic properties of the proposed method are evaluated through numerical experiments on datasets of varying sizes and structures, including grayscale and color images, text, and binary files. The effectiveness of the algorithm is assessed using entropy analysis, histogram uniformity, and differential attack resistance metrics (NPCR and UACI). Experimental results demonstrate that the entropy of encrypted data approaches the theoretical maximum (≈8 bits per byte), indicating near-ideal randomness and elimination of statistical dependencies. The proposed hybrid approach significantly enhances resistance to statistical and differential cryptanalysis compared to purely chaotic or classical cryptographic schemes. The results confirm the feasibility and effectiveness of integrating chaotic transformations with standardized cryptographic algorithms for secure multimedia data protection in modern information systems.