Ok, this may not be the answer you're looking for, but my advice would be to use standard tools that come with the operating system. For example, MS-DOS was released about 40 years ago and supported the concept of a file. Today, its successor, Microsoft Windows supports too the concept of a file. I bet whatever operating system you're using in 50 years time will too support the concept of a file.

So, shoot in JPG+RAW and store the files that your camera gives you. If you want to store custom metadata, don't modify the JPG+RAW files but instead create a TXT file with the same name and put your custom metadata there. Text files stored in US-ASCII format have been supported since 1969. If you're very adventurous, you may use UTF-8 which is a superset of US-ASCII and unlikely to vanish. Avoid ISO-8859-1; it will be replaced by UTF-8. Avoid any "rich text" formats; they will become outdated (except perhaps HTML, but it too is a fast-moving target).

JPG is a format that will surely be supported in 50 years. There are some improvements like JPEG2000 but they have not gained major acceptance because JPG is good enough and compatibility is more important than squeezing the last 10-20% of compression / image quality.

Camera RAW formats aren't strictly speaking open source, but however, open source utilities exist to read files in these formats. If the RAW becomes unreadable after 50 years due to nonexistent software, at least you have the JPG.

The main concern I have is how to store the bits in a way that lasts for 50 years. I would choose Amazon AWS S3 (approximately $0.0125 per gigabyte per month for infrequently accessed data) and store an offline backup copy on a hard disk or SSD. You may occasionally need to switch to a newer hard disk / SSD, and if Amazon S3 is shut down, switch to another cloud provider. Amazon S3 has a 13 year history so the chances of it lasting for 50 years are somewhat better than some random free cloud provider.