The environment light is changing based on daytime and night. Is it possible to use deep learning in this situation?
Mech-Mind products can resist to some extent the impact of changes in ambient lighting conditions. When there is a moderate variation in ambient light, the deep learning functionalities can still work effectively. However, if the lighting changes are too severe, it can still affect the performance of the vision software.
Severe changes in ambient lighting typically refer to:
Direct sunlight - Normal lighting - No/low lighting
Wide variation in brightness of artificial light sources
Changes in color temperature
Other significant lighting changes that can noticeably affect grayscale values in color images
If your working environment cannot avoid drastic natural light changes, I recommend designing appropriate shading or supplementary lighting solutions based on your project requirements. For shading solutions, you can refer to the following document: Fail Connection with Kuka Robot
Additionally, in Mech-Vision, the “image brightness and color balancer” step can be used to adjust the 2D image changes caused by minor variations in lighting conditions.
