Your Search Results

Use this resource - and many more! - in your textbook!

AcademicPub holds over eight million pieces of educational content for you to mix-and-match your way.

Experience the freedom of customizing your course pack with AcademicPub!
Not an educator but still interested in using this content? No problem! Visit our provider's page to contact the publisher and get permission directly.

A phenomenology-based approach to the automated recognition of materials in HYDICE imagery

By: Blake, P.; Carinhas, P.; Sharp, M.; Lundeen, T.; Hayashi, J.;

1998 / IEEE / 0-7803-4403-0


This item was taken from the IEEE Conference ' A phenomenology-based approach to the automated recognition of materials in HYDICE imagery ' In order to recognize an object's spectral signature and to identify that object, it is necessary to define the object's signature in contrast to relevant background materials, whether the object encompasses whole pixels or is part of a mixed pixel. The authors have developed an object recognition technique that performs automated materials recognition with respect to a spectral database. As part of the recognition process, the technique also orders the spectral database according to how similar a spectrum is to a given pixel. The technique performs this ordering at each step of an iterative demixing process, giving a hierarchical partitioning of the data with respect to a given pixel spectrum. The authors are using an ATR tool to understand database structure. The ATR tool is useful for this because it partitions the database as part of the process. The method operates on the whole pixel, reducing the pixel to mixed components if needed. The method also uses statistically significant thresholds to decide what is in the pixel. The authors show the application of this tool to determining spectral signatures for 5 selected objects in a HYDICE image, based on the relevant spectral database. They use an ATR tool, that called the residual correlation method (RCM). The RCM is based on an iterative solution to a set of linear mixing equations. The RCM can be used to recognize subpixel or whole pixel targets using multispectral or hyperspectral imagery. The foundation for the RCM is the linear mixing model (LMM). The LMM models each spectrum in an image as a linear combination of a set of contributing subpixel components.