Study on Stereo AI Based Zed-2i Camera

This article examines the possibility of successfully integrating the ZED-2i stereo camera with AI-enhanced technology. A sample of two different environments with different characteristics is assessed to determine the viability of implementing this in a real-world setting. The results show a significant difference in cloud detection between indoor and outdoor environments, with higher detail levels in point clouds. Essential factors include depth recognition, cloud clustering, and light conditions. Quality parameter settings were found to mitigate the adverse effects of localization and visual odometry inaccuracies. The findings from analyzing experiments display the importance of hardware and software configurations. Future research is required on the newer version of the camera with Jetson NX Orin implementation to explore and develop the utilization in specific sectors like agriculture and forestry.