Paul E. Anuta

Spatial Registration of Multispectral and Multitemporal Digital Imagery Using Fast Fourier Transform Techniques

A system for spatial registration of digitized multispectral and multitemporal imagery is described. Multispectral imagery can be obtained from sources such as multilens cameras, multichannel optical-mechanical line scanners, or multiple vidicon systems which employ filters or other spectral separation techniques to sense selected portions of the spectrum. Spatial registration is required so that multidimensional analysis can be performed on contextually similar image elements from different wavelength bands and at different times. The general registration problem is discussed first; then the fast Fourier transform (FFT) technique for cross correlation of misregistered imagery to determine spatial distances is discussed in detail. A method of achieving translational, rotational, and scaling corrections between images is described. Results of correlation analysis of multispectral scanner imagery and digitized satellite photography is presented. Use of the system for registration of multispectral airborne line-scanner imagery and space photography is described. Application of the techniques to preprocessing of earth resources satellite imagery from systems such as the earth-resources technology satellite (ERTS) scanner and vidicon system is discussed in conclusion.

Image Registration: Similarity Measure and Preprocessing Method Comparisons

An experimental comparison of several similarity measures and preprocessing techniques used for the registration of temporally differing images is carried out. It is found that preprocessing of the images via a gradient operator improves the registration performance. This is in agreement with a derived optimal processor (described in the Appendix) based upon image and temporal difference characteristics.

LANDSAT-4 MSS And Thematic Mapper Data Quality And Information Content Analysis

Landsat-4 Thematic Mapper and Multispectral Scanner data were analyzed to obtain information on data quality and information content. Geometric evaluations were performed to test band-to-band registration accuracy. Thematic Mapper overall system resolution was evaluated using scene objects which demonstrated sharp high contrast edge responses. Radiometric evaluation included detector relative calibration, effects of resampling, and coherent noise effects. Information content evaluation was carried out using clustering, principal components, transformed divergence separability measure, and numerous supervised classifiers on data from Iowa and Illinois. A detailed spectral class analysis (multispectral classification) was carried out on data from the Des Moines, Iowa area to compare the information content of the MSS and TM for a large number of scene classes.

Digital Registration of Multispectral Video Imagery

This paper describes an adaptive system for achieving spatial registration of digitized imagery. Multispectral imagery is obtained from sources such as multilens cameras, multichannel optical-mechanical line scanners, and multiple TV camera systems. The system converts the raster scan imagery to an array of binary numbers representing brightness at discrete points in the scene. An analysis method is described which estimates the difficulty of registering these multispectral digital pictures. This measure is used to control the registration algorithm to improve the rate of processing. Experiments are described which compare the performance of the system with and without adaptive control. Registration of test imagery in three categories is carried out and some results are presented.

Identification and area estimation of agricultural crops by computer classification of LANDSAT MSS data

Abstract LANDSAT Multispectral Scanner (MSS) data covering a three-county area in northern Illinois were classified using computer-aided techniques as corn, soybeans, or “other.” Recognition of test fields was 80% accurate. County estimates of the area of corn and soybeans agreed closely with those made by the USDA. Results of the use of a priori information in classification, techniques to produce unbiased area estimates, and the use of temporal and spatial features for classification are discussed. The extendability, variability, and size of training sets, wavelength band selection, and spectral characteristics of crops were also investigated.

Atmospheric effects on Landsat TM thermal IR data

The components of atmospherically attenuated target radiance and the path radiance emitted by the atmosphere are calculated to explain the fact that for certain meteorological conditions, properly calibrated thermal IR (infrared) data gathered from aircraft and spacecraft altitudes provide accurate temperature measurements of surface water bodies even when atmospheric corrections are not applied. Results show that although the 8-14- mu m atmospheric window is far from being transparent ( >

Crop surveys from multiband satellite photography using digital techniques

Abstract Multiband and Multibase photography of the Imperial Valley, California, taken during the Apollo 9 mission was analyzed using statistical multispectral pattern recognition techniques. With a scanning microdensitometer, 70-mm frames were converted into digital form and stored on magnetic tape. All analysis was conducted using a digital computer which enabled rapid automatic processing of any of over five million film density measurements. Automatic recognition of fields of barley, sugar beets, and alfalfa was attempted, and the recognition accuracies were 75% for barley, 59% for sugar beets, and 27% for alfalfa. Highly accurate detection of bare soil, salt flats, and water was achieved. Low percentage ground cover crops such as lettuce, onions, carrots, and cabbage were indistinguishable from bare soil. Results using multiband photography are compared with results using color separations of color infrared photography which covered approximately the same bands.

A Parametric Model for Multispectral Scanners

Efficient acquisition and utilization of remotely sensed data require an extensive a priori evaluation of the performance of the basic data collection unit, the multispectral scanner. The objective of the research described in this paper is the development of a parametric model to analytically evaluate the response of a multispectral scanner in any operational environment and to provide the necessary information in selecting a set of optimum parameters. In this paper the multispectral scanner spatial characteristics are represented by a linear shift-invariant multiple-port system where the N spectral bands comprise the input processes. The scanner characteristic function, the relationship governing the transformation of the input spatial and hence spectral correlation matrices through the systems, is developed. Specific cases for Gaussian point spread functions are examined. The integration of the scanner spatial model and a parameter classification error estimator provides the necessary technique to evaluate the performance of a multispectral scanner. A set of test statistics is specified and the corresponding output quantities computed by the characteristic function. Two sets of classification accuracies, one at the input and one at the output, are estimated. The scanners instantaneous field of view is changed and the variation of output classification performance computed.

Analytic signal representation of the synthetic seismogram of bright spots

Transformation of seismic signals into their analytic signal representation permits the unique separation of envelope, instantaneous phase, instantaneous frequency and apparent polarity. These parameters are useful in extracting the physical properties of a seismic signal and help in geophysical and geological interpretation. Deconvolution methods can improve the quality of the analytic seismic signal representation. From simulation studies it is found that time and space adaptive deconvolution significantly improves the quality of the analytic signal representation.

A two-dimensional filter design for isotropic reconstruction of track type airborne geophysical surveys

The problem of reconstruction of a continuous densely sampled uniform grid scalar surface from track-type geophysical surveys is discussed. Examples of track-type signal sources include earth gamma ray radiation and magnetic fields measured at low altitude (typically 500 ft.) along tracks spaced from a fraction to several miles apart. The signal model investigated assumes the geophysical surface consists of a narrow band isotropic stochastic signal process. The sampling process is characterized by a high sampling rate along track and a low sampling rate across track. A reconstruction filter approach is described which attempts to provide isotropic reconstruction to minimize the error in representation of scene features.