A Method for Generating Point Clouds From Monocular Depth Estimation Images

LiDAR sensors enable the acquisition of highprecision point clouds through the use of laser light pulses; however, their high cost is a significant limitation. In contrast, camera sensors offer a more economical and feasible alternative. This paper introduces a method for extracting point clouds akin to those produced by LiDAR, utilizing only a monocular camera sensor. The proposed methodology yields point clouds with acceptable precision and an uncertainty of approximately <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\pm 20 ~\text{cm}$</tex>, which can be further refined through precise camera parameter adjustments. The solution is compatible with monocular background estimation algorithms, camera calibration, and GPS integration, and can be easily adapted for use with other similar methods. Computational simulations confirmed the efficiency of the proposed method. Moreover, it is noted that the point cloud generated by LiDAR has less detail compared to the one generated by the proposed method, with elements at a very low additional cost.