I've noticed this too. As you have shown, you expect CPs to have greater error than GCP's but not as much as they seem to end up with in Context Capture. We use Pix4D and Context Capture.
↧