The rapid evolution of modern medical technology has propelled surgical robots to the forefront of surgical innovation. These robots offer unprecedented accuracy and safety in procedures, empowering surgeons with advanced technical capabilities to tackle complex medical challenges. In tumor treatment, surgical robots excel in precise tumor tissue resection, minimizing damage to healthy tissue and reducing surgical risks. However, accurately identifying tumor boundaries and depth remains challenging due to limitations of traditional palpation methods. This study aims to develop a deep learning model for real-time tumor boundary detection within surgical robot systems. By processing array tactile sensor image data, this model will enable precise boundary and depth detection, enhancing surgical robot recognition and operational capabilities. This advancement promises improved surgical success rates, patient survival rates, and reduced risks of complications, ultimately enhancing patient treatment outcomes and quality of life.