The corrected image of a .bmp has float type values while its range is between 0 and 255. So the pipeline.data['im_corrected'] returns a dark image (overflowed values).
It seems that in SilCamLoad() class, the '.bmp': skimage.io.imread does not rescale the image to float64 in the range of (0,1).
Suggestion: modifying the code by defining a lambda function for '.bmp' extension as:
'.bmp': lambda f: skimage.io.imread(f).astype(np.float64) / 255}
The corrected image of a .bmp has float type values while its range is between 0 and 255. So the pipeline.data['im_corrected'] returns a dark image (overflowed values).
It seems that in SilCamLoad() class, the '.bmp': skimage.io.imread does not rescale the image to float64 in the range of (0,1). Suggestion: modifying the code by defining a lambda function for '.bmp' extension as: '.bmp': lambda f: skimage.io.imread(f).astype(np.float64) / 255}