I would like to know how the image match module works. I have not found any documentation in the directory. Is this an old module that is no longer supported? When I train a folder of images, I get a recognition of 95%, but the preview is blank. I can not find any variables to say what image it has recognized.
I am trying to stitch a series of images into a panorama. I would like to read from a file of sequential JPG images and determine when the camera angle has made an entire revolution. The idea would be to train image match with the first photo, and then read each image until it finds one with the same (or close to the same) view. I have tried object recognition with the whole image, but that was not reliable. I am open to ideas.
Thank you,
Lane
|
|