后端优化的前提是前端位姿图的正确构建，图构建的准确性直接影响到后端优化

The premise of the back-end optimization is the correct construction of the front-end pose map, and the accuracy of the map construction directly affects the results of the back-end optimization. Wrong nodes and edges will cause the optimization result to deviate from the true value, resulting in an error in the final pose, which will affect the construction of the map. The research object of this paper is mainly 2D lidar and RGB-D camera. The following data fusion of 2D lidar and RGB-D camera for front-end data registration and loopback detection is used to improve the accuracy of pose image.

The premise of back-end optimization is the correct construction of the front-end position map, and the accuracy of the graph build directly affects the results of the back-end optimization. The wrong nodes and edges can cause the optimized result to deviate from the true value, causing the last position to go wrong, which affects the construction of the map. The main objects studied in this paper are 2D lidar and RGB-D cameras, and the data of 2D lidar and RGB-D cameras is combined for front-end data registration and loop detection to improve the accuracy of bitmap.

The premise of the back-end optimization is the correct construction of the front-end pose map, and the accuracy of the map construction directly affects the results of the back-end optimization. Wrong nodes and edges will cause the optimization results to deviate from the true value, resulting in the final pose error, which will affect the map construction. The research object of this paper is mainly 2D lidar and rgb-d camera. According to the front-end data registration and loop detection, the data of 2D lidar and rgb-d camera are fused to improve the accuracy of pose map.<br>