output metric depth compare with LiDAR data

[pxrw]

I employed the metric_vkitti model to generate metric depth data from the authentic Kitti dataset, subsequently reading LiDAR data for comparison. I observed significant discrepancies between the generated depth map results and the LiDAR data. For instance, when using image_2/000000.png as input, the generated depth yielded: depth max: 32.82895278930664, depth min: 6.090167999267578
gt max: 76.82421875, gt min: 4.22265625. Have I employed an incorrect methodology?

[Yunxiang-Wang]

Could it be that the standardized statistical data is incorrect?