Especially for Laser camera. What 2D parameter should be used to gather high quality image for deep learning?
You can check the high qualkity image in here: https://docs.mech-mind.net/en/dlk-software-manual/2.4.2/algorithm-modules/instance-segmentation-improve-model-accuracy.html
Furthermore, you can refer to this collection of parameters for making the corresponding modifications.https://community.mech-mind.com/t/posts-collection-guides-to-mech-eye-industrial-3d-cameras/531
Finally, the training images should cover as many different scenarios as possible, and the images used for actual prediction should be consistent.