Here’s an exciting experiment that tries to evaluate the high quality of a linear rail by making use of a variety of visual odometry, completed by mounting a digital camera on the rail and analyzing the movie with open up-resource computer software normally applied to stabilize shaky online video footage. No linear rail is perfect, and it must be possible to measure the diploma of imperfection by recording movie footage though the camera moves down the length of the rail, and examining the outcome. Imperfections in the rail ought to trigger the online video to sway a proportional total, which would permit one to characterize the rail’s good quality.
To exam this idea, [Saulius] hooked up a significant-definition digicam to a linear rail, pointed the camera to a high-contrast textured sample (generating the resulting video clip easier to review), and recorded video even though moving the camera throughout the rail at a set speed. The ensuing movie gets fed into the Deshaker plugin for VirtualDub, of which the important part is the deshaker.log file, which is made up of X, Y, rotate, and zoom correction values needed to stabilize the online video. [Saulius] utilized these values to generate a graph characterizing the linear rail’s top quality.
It is a intelligent proof of notion, specially in how it uses no specific instruments and leverages a video stabilizing algorithm in an unconventional way. Even so, the effects aren’t just effortless to turn into concrete, genuine-entire world measurements. Turning picture results into micrometers is a make a difference of counting pixels, and for this task video clip stabilizing is an imperfect device, due to the fact the algorithm prioritizes visual results instead of complete measurements. Nonetheless, it is an appealing experiment, and completely capable of measuring rail good quality in a relative feeling. Cannot assistance but be a little bit curious about how it would profile a little something like these cardboard CNC modules.
:quality(70)/cloudfront-us-east-1.images.arcpublishing.com/mco/KNNRUIN2EZBKDKFXGXCZNQKHCU.jpeg)
