Navigation rendering model with mapping identification, path planning and deviation correction functions

The method is based on a navigation rendering model with functions of identifying and mapping, planning a path and correcting deviation. The method comprises the following specific steps: according to a positioning position and a pre-collected image, performing three-dimensional rendering reconstruction by using a neural radiation field NeRF and a Nice-SLAM to realize indoor precise positioning; a YOLOv8n + DeepSORT algorithm is used to realize video target identification and correlation tracking, and deviation correction is realized; semantic information is added to a map through video understanding, a semantic map is achieved, and all planning results are rendered on a 3D navigation interface; and selecting an optimal path based on an improved DQN algorithm fusing a deep neural network and a reinforcement learning method. The system integrates sensing, positioning, data updating and path planning methods so as to provide more accurate, real-time and adaptable navigation experience, and aims to improve the p