This paper presents an integrated multi-sensor fusion approach for improving Simultaneous Localization and Mapping (SLAM) techniques, particularly for 3D reconstruction in dynamic environments. The proposed framework combines LiDAR-inertial odometry (LIO), visual-inertial odometry (VIO), and advanced Inertial Measurement Unit (IMU) preintegration methods to enhance the robustness and accuracy of SLAM. The integration of these techniques allows for the creation of detailed and accurate 3D maps, while also incorporating an object-detection network to identify and exclude dynamic obstacles such as pedestrians and vehicles, ensuring the integrity and precision of environmental mapping. The study addresses the challenges of mapping in environments with variable elements by leveraging the strengths of multiple sensor modalities. LIO provides high-accuracy spatial data from LiDAR, while VIO offers real-time visual data for feature extraction and tracking. The IMU preintegration method enhances the reliability of state estimation by aggregating inertial measurements over a fixed time window, reducing computational complexity and improving resilience against noise. The object-detection network, based on YOLOv5, enables efficient and accurate removal of dynamic obstacles, further improving the system's ability to navigate and map complex environments. The effectiveness of the proposed methods is validated through experimental evaluation, demonstrating their capability to produce more reliable and precise maps in dynamic settings. The integration of LIO and VIO within a unified framework allows for a more detailed and accurate representation of the mapped environments, providing a robust solution for SLAM in challenging and dynamic scenarios. The results indicate improvements in autonomous navigation and mapping, offering a practical solution for SLAM in environments with variable elements.This paper presents an integrated multi-sensor fusion approach for improving Simultaneous Localization and Mapping (SLAM) techniques, particularly for 3D reconstruction in dynamic environments. The proposed framework combines LiDAR-inertial odometry (LIO), visual-inertial odometry (VIO), and advanced Inertial Measurement Unit (IMU) preintegration methods to enhance the robustness and accuracy of SLAM. The integration of these techniques allows for the creation of detailed and accurate 3D maps, while also incorporating an object-detection network to identify and exclude dynamic obstacles such as pedestrians and vehicles, ensuring the integrity and precision of environmental mapping. The study addresses the challenges of mapping in environments with variable elements by leveraging the strengths of multiple sensor modalities. LIO provides high-accuracy spatial data from LiDAR, while VIO offers real-time visual data for feature extraction and tracking. The IMU preintegration method enhances the reliability of state estimation by aggregating inertial measurements over a fixed time window, reducing computational complexity and improving resilience against noise. The object-detection network, based on YOLOv5, enables efficient and accurate removal of dynamic obstacles, further improving the system's ability to navigate and map complex environments. The effectiveness of the proposed methods is validated through experimental evaluation, demonstrating their capability to produce more reliable and precise maps in dynamic settings. The integration of LIO and VIO within a unified framework allows for a more detailed and accurate representation of the mapped environments, providing a robust solution for SLAM in challenging and dynamic scenarios. The results indicate improvements in autonomous navigation and mapping, offering a practical solution for SLAM in environments with variable elements.