I am exploring the feasibility of creating an open data set of imagery (and potentially other data) that can be used to train algos for search and rescue operations (and potentially wider humanitarian applications). FYI, this is some of the work I've been leading on: https://www.wired.co.uk/article/rnli-coastguard-drones-search-rescue-missions.
Data standards - how might the data be collected and codified in such a way as to make it appropriate for exploitation for training?
Chip detection - how might we help in the training of algorithms, using crowdsourcing etc?
How might organisations easily contribute data and how might it be accessed? What are the conditions of that access? Humanitarian use only?
Some project benefits:
- Improved detection probability
- Improved accuracy and lowered false alarm rates
- Improved return on investment for trials
- Lower cost of data acquisition
- Reduced time lag to acquiring data
- Access to multiple imagery capture devices
- Enhanced, more cost effective sensor payloads
- UAS enthusiasts can contribute data to humanitarian cause
- Leverage the crowd to identify objects of interest, allowing people to contribute skills to a humanitarian cause.
Is an open project like this something that would excite you? Is it duplicating something that is already out there?
Thoughts / rock throwing / feedback encouraged.