In this paper, we introduce a novel target-less LiDAR-Camera Auto Calibration (TLAC) method, leveraging foundation models to achieve high-precision alignment without the need for physical calibration targets. Our approach utilizes advanced processing techniques for both point cloud data (PCD) and images, incorporating voxel downsampling, normal estimation, and clustering for PCDs, along with semantic segmentation and depth estimation for images. We validate our method using the widely recognized KITTI odometry benchmark dataset, specifically focusing on sequence 00 for a comparative analysis with existing techniques. Our experimental results demonstrate significant improvements in calibration accuracy, showcasing the potential of our method to facilitate enhanced sensor fusion for autonomous vehicles and robotics. This study not only advances the field of sensor calibration but also highlights the importance of integrating diverse data processing techniques for improved environmental perception.