Build a prefect request /session to avoid 403 error for my scraper on a specific website -- 2

This website has CloudFlare protection. When too many requests are sent, google recaptcha will ask verification. When using proxy in my browser, I can bypass this but the weird thing is no matter how I construct my header with proxy, my request will always get 403 forbidden. There must be something wrong with my request as the proxy is verified by using selenium or just use browser. I need a scraper expert to help me diagnose or build a perfect request then I can continue my scraping. You should us python3, request or session.