Showing papers by "David A. Landgrebe published in 1987"

An unsupervised feature extraction method for high dimensional image data compaction

Abstract: A new on-line unsupervised feature extraction method for high-dimensional remotely sensed image data compaction is presented. This method can be utilized to solve the problem of data redundancy in scene representation by satellite-borne high resolution multispectral sensors. The algorithm first partitions the observation space into an exhaustive set of disjoint objects. Then, pixels that belong to an object are characterized by an object feature. Finally, the set of object features is used for data transmission and classification. The example results show that the performance with the compacted features provides a slight improvement in classification accuracy instead of any degradation. Also, the information extraction method does not need to be preceded by a data decompaction.

Stochastic Model Utilizing Spectral and Spatial Characteristics

Abstract: In remote sensing, because of physical properties of targets, sensor pixels in spatial proximity to one another are class conditionally correlated Our main objective is to exploit this spatial correlation Therefore, a two-dimensional causal first order Markov model was used to extract the spatial and spectral information and, based upon it, new object classifiers with improved performance were developed First, the minimum distance (MT) and the maximum likelihood (ML) object classifiers are discussed Then, based on the proposed model, these two classifiers are modified, and a linear object classifier is introduced Finally, experimental results are presented

Spectral feature design for data compression in high dimensional multispectral data

Abstract: Data transmission loads of high dimensional remote sensor systems can be greatly reduced by applying generalized Karhunen-Loeve transform as a feature design technique. Two spectral feature design approaches based upon the generalized K-L transform are developed to compress information effectively. Six sets of field data from Kansas and North Dakota on three different dates each are used to test the methods. Spatially, temporally and spatially/temporally combined data sets are formed in this paper to test the robustness property of the schemes. The probability of correct classification using Landsat MSS, Thematic Mapper bands and the proposed bands are found and compared. The comparison shows that the results are improved by the proposed methods, and they appear to be satisfactorily robust. The overall data compression ratio in this paper is about 100/16, i.e., about 6 to 1 with no loss in classification accuracy.