在计算机视觉的研究中, 同时定位和地图构建(SLAM) 是一类极具挑战性的复杂问题. ACP 理论为复杂系统建模与调控提出了一条新的有效途径. 在提出的平行感知的基本框架及其关键技术中,ACP 理论被开创性地引入SLAM 算法中. 平行感知利用各种人工场景数据和采集到的实际场景数据, 通过计算实验进行视觉SLAM 模型及参数的训练与评估, 然后借助平行执行来在线优化视觉系统, 实现对复杂环境的智能感知与理解, 最终建立了一个从底层视觉算法到高层决策与分析的全新理论.; In the computer vision research, simultaneous localization and mapping (SLAM) is one of the most challenging problems. The ACP theory provides a new and e ective solution for modeling and control of complex systems. In the proposed parallel perception and its basic framework and key techniques, the ACP theory is innovatively introduced into the SLAM algorithm. For parallel perception, artificial scenes are used to represent complex real scene, computational experiments are used to train and evaluate models and parameters of visual SLAM algorithms, and parallel execution aims to continuously optimize the vision system and achieve the intelligent perception and understanding of complex systems. As a result, a novel and new perception theory is constructed to integrate low-level vision algorithms and high-level decision and analysis.