Using BRISK points on a low contrast image

I'm doing a project where I need to match a template image of a stitch to an image of knitwear given to the program. I have been given the added difficulty of not being able to take direct crops from the image provided and instead have made a database of images to try to match. I am ultimately trying to reverse engineer a piece of knitwear into a pattern of knits and purls (the two different types of stich in knitwear).

Before anything else, all images are turned to greyscale and have imadjust applied to increase their contrast.

My issues arise initially because I'm trying to use BRISK points to match my templates. I have the program look at the provided knitwear image in sections that are the same size as the templates. It does the comparison and throws no errors. However, most of the templates and sections do not detect any BRISK points and, when they do, they do not match any points.

Any suggestions of how I could fix this or of an alternate method I could use to try to get matches, it would be much appreciated.

I have attached some sample images of the templates and the kind of image I would be reverse engineering.