Constrained subpixel target detection for hyperspectral imagery

The detection in remotely sensed images can be conducted spatially, spectrally or both. The difficulty of detecting targets in remotely sensed images with spatial image analysis arises from the fact that the ground sampling distance is generally larger than the size of targets of interest in which case targets are embedded in a single pixel and cannot be detected spatially. Under this circumstance target detection must be carried out at subpixel level and spectral analysis offers a valuable alternative. This paper compares two constrained approaches for subpixel detection of targets in remote sensing images. One is a target abundance-constrained approach, referred to as the nonnegatively constrained least squares (NCLS) method. It is a constrained least squares linear spectral mixture analysis method which implements a nonnegativity constraint on the abundance fractions of targets of interest. A common drawback of linear spectral mixture analysis based methods is the requirement for prior knowledge of the endmembers present in an image scene. In order to mitigate this drawback, the NCLS method is extended to create an unsupervised approach, referred to as the unsupervised nonnegatively constrained least squares (UNCLS) method. This unsupervised method can be implemented with only partial or no prior knowledge of targets present in the image scene. The second approach is a target signature-constrained method, called the constrained energy minimization (CEM) method. It constrains the desired target signature with a specific gain while minimizing effects caused by other unknown signatures. Data from the HY perspectral Digital Imagery Collection Experiment (HYDICE) sensor are used to compare the performance of these methods.