Shadow detection is an important step in the shadow phenomenon,and the total variational model can be used in shadow detection.This paper improves the TV model,and puts forward a shadow detection algorithm based on unconstrained total variation.After the experiments with several real images and statistical analysis,it shows that the unconstrained total variational shadow detection is valid for single iterative images.Total variation model can be used in shadow detection,which is an important step before shadow processing.An unconstrained total variation model is proposed to detect shadow,based on the improvement of variation model.Experiment and statistical analy-sis show that the model can detect shadow of single image under appropriate iteration conditon.