An unsupervised hierarchical clustering image segmentation and an adaptive image reconstruction system for remote sensing

The evolution of technology is radically affecting the quantity and quality of data collected in many scientific disciplines. This is especially true in the disciplines which contribute to the earth-science data information system due to innovations in "digital remote sensing." Remote sensing is a general term which includes aerial surveys and sonar and radar mapping, but which is primarily becoming applied to digital image data from satellites.; Recently, there has been increasing interest in the use of statistical methods to analyze the highly structured data of digital images. Statistical approaches in image processing have close ties with multivariate analysis and decision theory. This study is concerned with statistically-based image analysis, principally for applications in remote sensing.; A multistage algorithm which makes use of spatial contextual information in a hierarchical clustering procedure has been developed for unsupervised image segmentation. A Markov random field model is employed to enforce local spatial smoothness, while the maximum entropy principle is utilized to quantify global smoothness in the image. A multi-window approach implemented in a pyramid-like data structure which uses a so-called boundary blocking operation is employed to increase computational efficiency. The Schwarz information criterion is suggested as a means of selecting the level in the clustering hierarchy which corresponds to the optimal state.; An adaptive reconstruction system has also been developed to analyze sequential images observed at regular time intervals. A least-squares linear predictor with escalator structure has been implemented in the new system. Using the predictor, estimates of missing data or bad (possible cloud covered) data and the spatial parameters at a specified time can be estimated based on previous history. This algorithm recovers from observations which are contaminated due to blurring and/or correlated noise using temporally adapted spatial parameters. The reconstruction system can be used either individually for improving the pictorial information in the image or as a preprocessor for the image segmentation algorithm.; The segmentation procedure and the reconstruction system have been evaluated extensively using simulated data and applied to remotely sensed images from NOAA-n satellites.