Now to detect a print, high-speed through 30/SEC, length is 1 m, 0.5 m wide,
Plans to use the photo taking pictures, and then use the AI algorithms are classified,
Array camera with lines, to generate the smallest photos of 20 m, more than 40 m, for AI, pure waste, not so high resolution,
But ordinary camera, testing ghosting, any Suggestions?

CodePudding user response:

Don't want to, too fast, not achieve them

CodePudding user response:

Under the shutter speed to 1/30

CodePudding user response:

I think you can be solved from two ideas, one is that by selecting the low resolution of linear array camera to reduce the image size, another idea is to use the conventional algorithm of the image of the original image processing import ai algorithms, each image set aside 300 ms processing time required is not high,

CodePudding user response:

Thank you, ready to try the low resolution of linear array,

CodePudding user response:

Can first send a sample to the right size, and then throw it to the AI, three threads

CodePudding user response:

30 pieces per second and it is a larger version of such a high resolution, is also a challenge in modern industry, the need to reduce the resolution, if use within DNN current work force is not so fast,