Your leaf segmentation is lousy because you started with a bad picture. Take the picture again with a black velvet background behind the leaf. That beige background you have is not allowing good segmentation. With a black background it would be SO much easier. Also, put the leaf under a thin sheet of glass to reduce shadows, though that should not be a problem if you use the black background. Finally you need to make sure you have good control over the light. Don't use a flash since they vary too much in light output from snapshot to snapshot. Use broad lighting, not a point source. Measure the light illuminance with a luxmeter to make sure it's consistent from image to image. If you really want to be accurate, you'd have an X-rite ColorChecker chart in each image and calibrate the image to LAB colorspace (like my icon to the left).
For color classification, you might want to take a look at the demo in the Image Processing Toolbox: https://www.mathworks.com/help/images/examples/color-based-segmentation-using-the-l-a-b-color-space.html which does it in LAB colorspace though not calibrated lab colorspace (which you'd need the ColorChecker chart and my attached directions for).
