The Internet of Things (IoT) involves the transmission of vast amounts of data in every single second, posing critical security challenges that demand robust solutions to ensure confidentiality and integrity. This study presents a novel hybrid security framework Lorenz-Rössler-Elliptic Curve Cryptography (LRE) that integrates with an innovative XOR-Matrix-based encoding steganography technique to enhance data security in interconnected IoT devices. LRE encrypts sensitive data from IoT devices, ensuring strong cryptographic protection, while the encrypted data is concealed within images using the steganographic method for improved data concealment. An Adaptive Firefly Optimization technique is employed to select optimal embedding blocks in the image, thereby maximizing embedding efficiency and security. Additionally, we propose an authentication protocol that enables IoT device authentication before data exchange in IoT networks. The proposed protocol ensures secure communication between IoT devices using lightweight session keys and Elliptic Curve Cryptography (ECC), ensuring low computational overhead. It employs implicit key-based authentication with global IDs and nonces to enable efficient mutual authentication and prevent security breaches. The framework's performance is rigorously evaluated using various metrics and compared with existing approaches, showcasing superior encryption strength, embedding robustness, and computational efficiency. The proposed protocol also ensures secure and reliable IoT communication by resisting DoS, replay, and man-in-the-middle attacks through efficient authentication and session key management. Precise data extraction and decryption further validate the framework's reliability for secure data transmission. By addressing key vulnerabilities and introducing an advanced optimization-based steganographic approach, this hybrid framework offers a scalable and effective solution to enhance data security in resource-constrained IoT environments.