Detection of leafy spurge using hyper-spectral-spatial-temporal imagery

We examined the ability of hyper spectral (80 bands), spatial (0.3 m), and temporal (10 dates during the growing season) imagery to detect leafy spurge infestations and classify plant densities. Random forest classification was used for all analyses. Single date classifications were similar to the best classifications in other studies (73% to 90% overall accuracies), although with greater distinction of densities. Mulitdate classification achieved 97% accuracy over four density classes.