Yes, it can. GluonCV: a Deep Learning Toolkit for Computer Vision does the job. One just need to go with an instance segmentation: Segmentation — gluoncv 0.11.0 documentation.
Then from there they can display the data however they feel valid. Input should be a video file (can be live but post firing is easier to process) then the instance segmentation can bookmark the timestamp and an image composition can display the target shots in order.
Personally I may throw this up on the community grid to see if one can capture target data from a drone firing an IR beam of non-lethal origin towards a receiver / target plain. Been meaning to figure that one out for a few months now and would be a great way to help train some ai models.