Firstly, a good idea is to roughly estimate reputation through citation metrics. Determine 3-5 most reputable journals in your field of interest (typically that would include Nature & PhysRev Letters for physics). To find these journals, go through some major lab publication lists to compile the names, and then look up citation metrics for them. Look through your shortlist of labs & professors and you'll see some of them would get 2-3 papers in these per year, while others maybe can get there once per several years. Also you can directly look up citation metrics for professors on Google Scholar or similar services.
Now, while reputation and citation metrics are important, it tells little about their ability to handle PhD students. To get an idea about that, for each lab compile a list of recently graduated students and try to find where they are now. Are they at major world-known universities/labs? What are their positions - are they jumping from one postdoc to another for years, or do they have a more respectable position after just one or two postdocs? What did they publish while studying, are their papers in reputable journals, are they well-cited? And not the least important, how many years did it take them to graduate? If an average student spends 6+ years on their PhD, it's probably a red flag.
When researching publications, discard any conference & poster presentations - students can get plenty of those with very little effort. Look only for legitimate publications, and ensure they are well-cited (ignoring self-citations).
Typically, reputation of labs is well-correlated with university reputation. Of course, there are some dying labs in respected unis, there are some great labs in otherwise mediocre ones. But start from top universities. After reading through publications and citations for a few hours you'll pick up on familiar names and get the general idea about the major influencers.