Detections on binary files by assumingly interpreting it as text

### Description

When scanning some larger projects we noticed that unextracted archives will be scanned by Scancode for licenses, leading e.g. to detections on an .tar archive based on it containing a packaged LICENSE-file. We are assuming that this impacts all binary files.
This would mean that interpreting any binary as text could lead to wrong detections depending on what is decoded from the binary.

### How To Reproduce

Scan a binary file with Scancode.