This paper presents a constrained energy minimization method of sub-pixel target detection for hyper-spectral imageries based on minimum noise fraction (MNF-CEM). In this method

minimum noise fraction (MNF) reduces the dimension of the hyper-spectral imageries and separates the noise from the hyper-spectral imageries. Then constrained energy minimization method (CEM) is used to detect the sub-pixel targets in the imageries of low dimension to void the computations of ill-conditional matrix inverse and virtual dimension of the hyper-spectral imageries. Experiments show that this method can restrain the influence of noise and is an effective method of sub-pixel target detection for hyper-spectral imageries.