In the multi-sensor data processing of the Cartographer algorithm, outliers and noise in the point cloud affect the accuracy of point cloud matching, and the accuracy of the pose fusion algorithm is not high. Therefore, an improved Cartographer algorithm based on the hybrid filtering algorithm and the pose fusion algorithm with velocity integral was proposed. First, the selected point measurement was improved to optimize the re-sampling process of the voxel filtering algorithm and improve the filtering efficiency. By introducing straight-through filtering and radius filtering, a hybrid filtering algorithm was proposed to improve the quality of the point cloud. Then, in the algorithm of fusing the observation pose, odometry data and inertial measurement unit data, the pose fusion algorithm with speed integration was introduced to improve the accuracy of point cloud matching. Finally, in the test experiment using the data set to verify the loop detection performance and the localization accuracy of the algorithm, the results show that compared with the cartographer algorithm and the A-LOAM algorithm, the map constructed by the improved algorithm is more accurate and the trajectory error is smaller. Therefore, the algorithm proposed in this paper is feasible and effective for improving localization accuracy and mapping quality.