Quantification of forest canopy changes caused by spruce budworm defoliation using digital hemispherical imagery

•Hemispherical images were used to predict annual spruce budworm defoliation.•Images quantified gap factor changes (GFC) from before and after defoliation.•GFC, balsam fir basal area %, and insecticide spraying served as model inputs.•GLM and Random Forests models predicted defoliation with RMSE of 14–24%.•Models including spraying improved predictions and is recommended if available.