The picture per day part is really cool - instead of having to look for a life raft/debris, we can look for patterns of abnormality across time. Given how large the ocean is, we have a good sense of the 'normal' distribution of images every day, so maybe we can find patterns like consistently "abnormal" images along a wind/current vector over 7 days. Individual debris may not be visible, but fields of debris may have some characteristics we can recognize.