There are two approaches, one potentially of zero additional cost.
If you are going to be using a colour camera, a typical DSLR for instance, it already comes with a set of three filters — one each for red, green and blue. None of these are standard photometric filters but it is possible to make usefully accurate estimates of what the intensity would have been in standard bands such as Johnson-V. The process takes far too many words to explain here but comprehensive instructions are available on-line. I don’t have URLs immediately to hand but can dig out the references later if you want them (and if you haven’t found them independently).
If you have a mono camera and/or wish to purchase a filter to begin your research-quality observation program, I would recommend buying a Johnson-V filter. They are available from a number of suppliers but please be careful to buy a Johnson-V photometric filter. Supplier’s product listings can be confusing at times.
In the case of DSLR (and other colour) cameras, always download your images in RAW format. Conversion to JPG or whatever will destroy the accuracy of your data. If you are using a filter as well, post-process the images first to separate out the RGB components and then sum or average (not median) combine them to produce a grey-scale image for subsequent analysis.
If you have a colour camera then I would personally use it as-is to learn the basics of the trade before spending more money. But, then, I’m a cheapskate. You are going to have to educate yourself anyway …