LARS Technical Reports Copyright (c) 2013 Purdue University All rights reserved. http://docs.lib.purdue.edu/larstech Recent documents in LARS Technical Reports en-us Wed, 13 Feb 2013 09:47:01 PST 3600 An Evaluation of Machine Processing Techniques of ERTS-1 Data for User Applications http://docs.lib.purdue.edu/larstech/141 http://docs.lib.purdue.edu/larstech/141 Fri, 23 Oct 2009 15:56:19 PDT For some years research has been underway to learn how to use computing machines for the processing and analysis of remotely sensed data. The motivation of this work is not to produce an automatic analysis system, but to utilize computing machinery for the repetitive and routine aspects of the total analysis and processing problem in order to minimize the costs and increase the throughput rate.

]]> David Landgrebe Techniques for Computer-Aided Analysis of ERTS-1 Data, Useful in Geologic, Forest and Water Resource Surveys http://docs.lib.purdue.edu/larstech/139 http://docs.lib.purdue.edu/larstech/139 Fri, 23 Oct 2009 15:56:18 PDT Forestry, geology. and water resource applications were the focus of this study, which involved the use of computer-implemented pattern-recognition techniques to analyze ERTS-1 data. The results have proven the value of computer-aided analysis techniques, even in areas of mountainous terrain.

Several analysis capabilities have been developed during these ERTS-1 investigations. A procedure to rotate, deskew, and geometrically scale the MSS data results in 1:24,000 scale printouts that can be directly overlayed on 7 1/2 minute U.S.G.S. topographic maps. Several scales of computer-enhanced "false color-infrared" composites of MSS data can be obtained from a digital display unit, and emphasize the tremendous detail present in the ERTS-1 data. A grid can also be superimposed on the displayed data to aid in specifying areas of interest, such as avalanche tracks or areas of burned-over timberland. Temporal overlays of six sets of data have allowed both qualitative and quantitative analysis of changes in the areal extent of the snowpack.

Computer-aided analysis of the data allows one to obtain both cover-type maps and tables showing acreage of the various cover types, even for areas having irregular boundaries, such as individual watersheds. Spectral analysis of snow and clouds, water and shadow areas, and forest cover of varying overstory density have revealed several important results.

]]> Roger M. Hoffer Mapping Soils, Crops, and Rangelands by Machine Analysis of Multi-Temporal ERTS-1 Data http://docs.lib.purdue.edu/larstech/140 http://docs.lib.purdue.edu/larstech/140 Fri, 23 Oct 2009 15:56:18 PDT ERTS-1 data, obtained during the period 25 August 1972 to 5 September 1973 over a range of test sites in the Central United States, have been used for identifying and mapping differences in soil patterns, species and conditions of cultivated crops, and conditions of rangelands. Multispectral scanner data from multiple ERTS passes over certain test sites have provided the opportunity to study temporal changes in the scene.

Geometric correction was performed on the digital data for several dates and for several test sites. This made much easier the task of locating specific data points and of comparing the analytical results with other maps and data sources.

Multispectral classifications delineating soils boundaries in different test sites compare well with existing soil association maps prepared by conventional means.

Spectral analysis of ERTS data was used to identify, map, and make areal measurements of wheat in western Kansas.

Multispectral analysis of ERTS-1 data provided patterns in rangelands which can be related to soils differences, range management practices, and the extent of infestation of grasslands by mesquite (Prosopis fuliflora) and juniper (Juniperus spp.)

]]> M. F. Baumgardner et al. Annual Technical Summary Report http://docs.lib.purdue.edu/larstech/137 http://docs.lib.purdue.edu/larstech/137 Fri, 23 Oct 2009 15:56:17 PDT This Final Technical Summary summarizes work done and results obtained in a broad set of remote sensing research studies. These include studies of spectral characteristics of crops and soils; the status of a field research data bank and a software data handling and analysis system for it; specification of a standardized multispectral field data acquisition system; studies of machine implemented training sample labeling methods; scene stratification and area estimation procedures for future crop inventory systems; a software system for studying the optimality and effectiveness of various sets of multispectral scanner parameters; the construction of multitype data sets involving the combination of Landsat data with synthetic aperature radar data and map-derived data; an assessment of the methods used by various countries of North and South America and Asia for acquiring, analyzing and reporting crop production statistics; and the status of a multilocation, multiterminal computer processing system for supporting remote sensing research. More complete reports on the studies are contained in the five volume final report of the contract and in various technical reports, which are listed in the technical summary.

]]> D. A. Landgrebe et al. Remote Sensing, Computers, and Land Use Planning http://docs.lib.purdue.edu/larstech/138 http://docs.lib.purdue.edu/larstech/138 Fri, 23 Oct 2009 15:56:17 PDT During the past decade, increased emphasis has been placed on land use planning at state, regional and national levels. A thorough knowledge of present land use is prerequisite to planning future land use. In other words land use inventory or classification is a primary stage in comprehensive planning at all levels.

]]> Harry C. Hitchcock Field Research on the Spectral Properties of Crops and Soils http://docs.lib.purdue.edu/larstech/135 http://docs.lib.purdue.edu/larstech/135 Fri, 23 Oct 2009 15:56:16 PDT This report describes the experiment design, data acquisition and preprocessing, data base management, analysis results and development of instrumentation for the AgRISTARS Supporting Research Project, Field Research task. It reports results of several investigations on the spectral reflectance of corn and soybean canopies as influenced by cultural practices, development stage and nitrogen nutrition. Results of analyses of the spectral properties of crop canopies as a function of canopy geometry, row orientation, sensor view angle and solar illumination angle are presented. The objectives, experiment designs and data acquired in 1980 for field research experiments are described. The development and performance characteristics of a prototype multiband radiometer, data logger, and aerial tower for field research are described.

]]> M. E. Bauer et al. Vol. II. Research in the Application of Spectral Data to Crop Identification and Assessment http://docs.lib.purdue.edu/larstech/136 http://docs.lib.purdue.edu/larstech/136 Fri, 23 Oct 2009 15:56:16 PDT Four investigations of the application of spectral data to crop identification and assessment are discussed in this volume. Part A discusses the development of spectromet crop development stage models. Several "state-of-the-art" development stage models for corn and soybeans were reviewed. One photothermal model and four thermal models were selected and evaluated.

Part B describes an investigation of spectral data as a source of information for crop yield models. Intercepted solar radiation and soil productivity are identified as factors related to yield which can be estimated from spectral data.

Several techniques for machine classification of remotely sensed data for crop inventory are evaluated in Part C. Early season estimation, training procedures, the relationship of scene characteristics to classification performance, and full-frame classification methods have been studied.

In part D, a task to determine the optimal level for combining area and yield estimates of corn and soybeans is discussed. The optimal level is assessed utilizing current technology: digital analysis of Landsat MSS data on sample segments to provide area estimates and regression models to provide yield estimates.

]]> C. S. T. Daughtry et al. Agricultural Scene Understanding http://docs.lib.purdue.edu/larstech/133 http://docs.lib.purdue.edu/larstech/133 Fri, 23 Oct 2009 15:56:15 PDT Results of four investigations, all related to agricultural remote sensing are described. The four tasks are: (A) Analysis of Agronomic and Spectral Data for Physical Understanding, (B) Field Measurements Data Management, (C) Multicrop supporting Field Research, and (D) Determining the Climatic and Genetic Effects on the Relationships Between Muitispectral Reflectance and Physical-Chemical Properties of Soils.

A. The Analysis of Agronomic-Spectral Data report describes the results of analyses of LACIE Field Research Data, including the relationships of agronomic and reflectance characteristics of wheat canopies, effect of cultural and environmental factors on reflectance properties of wheat, and discrimination of wheat and other crops as a function of wavelength band selection and acquisition date.

B. The Field Measurements Data Management report describes field research data base developed at LARS including the development of graphical and statistical analysis software, data processing software, and distribution of data.

C. The Multicrop Supporting Field Research report describes the measurements of spectral characteristics of corn and soybeans and development of a multispectral data acquisition system for field research.

D. The fourth report describes the objectives, experimental approach, and initial results of a study of the relationships between the reflectance and physical-chemical properties of over 400 different soils.

]]> M. E. Bauer et al. The Effects of the Physical and Chemical Properties of Soils on the Spectral Reflectance of Soils http://docs.lib.purdue.edu/larstech/134 http://docs.lib.purdue.edu/larstech/134 Fri, 23 Oct 2009 15:56:15 PDT Many attempts have been made to use multispectral data to map soils on the basis of soil color, texture, organic matter content, moisture content, and free iron oxides. Researchers have endeavored to differentiate soil series and coil types, using those parameters individually. Reflective properties have been attributed primarily to one or two of these parameters. Little has been revealed as to the nature of the natural interactions occurring between those parameters. Scientists have suggested that a more complete understanding of the radiation properties of specific soil constituents and their interactions is needed before one can optimize the use of multispectral separability in differentiating soil series and soil orders.

The objectives of this study are twofold: (1) To evaluate quantitatively the effects of organic matter, free iron oxides, texture, moisture content, and cation exchange capacity on the spectral reflectance of soils, and (2) to develop and test techniques for differentiating soil orders by computer analysis of multispectral data.

By collecting 71 soil samples of benchmark soils from the different climatic regions within the United States (having different vegetative cover types, parent materials, and geological history) and using the extended wavelength field spectroradiometer (Exotech Model 20B) to obtain reflectance values and curves for each sample, average curves were constructed for each soil order (excluding Oxisols and Histosols).

Multiple regression analyses were performed, using the spectral data as the dependent variables and physical - chemical properties as independent variables. The independent variables showing the highest correlation with the multispectral measurements were CEC and silt content. These results suggest that multispectral analysis may be a valuable tool for delineating and quantifying differences between soils.

]]> O. L. Montgomery et al. Moire Patterns and Two-Dimensional Aliasing in Line Scanner Data Acquisition Systems http://docs.lib.purdue.edu/larstech/131 http://docs.lib.purdue.edu/larstech/131 Fri, 23 Oct 2009 15:56:14 PDT The basic mechanism underlying the generation of Moire patterns in line scanner data acquisition systems is examined. A general expression is developed in tern of typical system parameters for the reproduced image of such systems and the interaction of the image spectrum; the raster frequency and digital sampling frequency of the A/D conversion process are discussed and examples given. System design requirements for avoiding Moire pattern generation and two-dimensional aliasing are discussed.

]]> C. D. McGillem et al. Conjugate Point Determination in Multitemporal Data Overlay http://docs.lib.purdue.edu/larstech/132 http://docs.lib.purdue.edu/larstech/132 Fri, 23 Oct 2009 15:56:14 PDT The machine processing of spatially variant multitemporal data such as imagery obtained at different times requires that these data be in geometrical registration such that the analysis processor may obtain the datum for a specified ground resolution element in each of the sets of imagery being utilized for analysis.

Misregistration between corresponding subsets of imagery contains both a displacement and a geometrical distortion component, and the affine transformation is postulated to characterize this misregistration between data subsets. Search techniques utilizing the moduli of the Fourier Transforms of these data are developed for estimating the coefficients of geometrical distortion components of this model.

Following the correction of these distortion components, the displacement is located by the crosscorrelation of a template obtained from one set of data, termed the reference, with the second, or background data. This template, derived for the optimum discrimination of the reference data embedded in the background, is determined by the solution of a system of equations involving the reference data and the covariance matrix of these data.

The derivation of the optimum filter includes constraints such that the maximum filter output, corresponding to the correct superposition of the reference template on the background data, is unity and the energy in the filter is finite. The filter obtained in this development is linear although it may involve a parameter requiring the solution of a nonlinear equation.

The performance of the crosscorrelation algorithm is evaluated using ideal data obtained by convolving an array of computer generated random numbers with a two-dimensional lowpass filter having a specified impulse response. The results obtained from these data generally substantiate the conclusions drawn from the analysis of this algorithm. The correlator output is then obtained for noise free and distortionless line scanner data. In these data the reference is selected as a subimage of the background data, and the data are selected to typify line scanner imagery. Multitemporal data are processed with the algorithms developed for the noise-free data to evaluate the applicability of this filter to the conjugate point problem.

It is demonstrated that the crosscorrelation of the template derived from the reference data will not yield useful results unless the geometrical correction of the data is implemented. The Fourier transform search techniques are used to estimate the distortion model coefficients, and a bilinear interpolation algorithm is utilized to correct the imagery. Results of the processor output using the corrected data are given. It is shown that the optimum filter yields a more discriminable peak of the correlation surface at the correct superposition of the reference template on the background than does the filter chosen as a subimage of the reference data itself.

]]> Richard A. Emmert et al. Pattern Recognition: A Basis for Remote Sensing Data Analysis http://docs.lib.purdue.edu/larstech/129 http://docs.lib.purdue.edu/larstech/129 Fri, 23 Oct 2009 15:56:13 PDT Pattern recognition plays a central role in numerically oriented remote sensing systems. It provides an automatic procedure for deciding to which class any given ground resolution element should be assigned. The assignment is made in such a manner that on the average correct classification is achieved.

This information note describes briefly the theoretical basis for the pattern-recognition-oriented algorithms used in LARSYS, the multispectral data analysis software system developed by the Laboratory for Applications of Remote Sensing (LARS).

]]> Philip H. Swain Utilizing Remote Multispectral Scanner Data and Computer Analysis Techniques http://docs.lib.purdue.edu/larstech/130 http://docs.lib.purdue.edu/larstech/130 Fri, 23 Oct 2009 15:56:13 PDT This research was designed to study the ability of present automatic computer analysis techniques with the use of multispectral scanner data to differentiate land use categories represented in a complex urban scene and in a selected flightline. An airborne multispectral scanner was used to collect the visible and reflective infrared data.

A small subdivision near Lafayette, Indiana was selected as the test site for the urban land use study. Multispectral scanner data were collected over the subdivision on May 1, 1970 from an altitude of 915 meters. The data were collected in twelve wavelength bands from 0.40 to 1.00 micrometers by the scanner.

The results indicated that computer analysis of multispectral data can be very accurate in classifying and estimating the natural and man-made materials that characterize land uses in an urban scene.

A 1.6 km. wide and 16 km. long flightline located in Sullivan County, Indiana, which represented most major land use categories, was selected for analysis. Multispectral scanner data were collected on three flights from an altitude of 1,500 meters. Energy in twelve wavelength bands from 0.46 to 11.70 micrometers was recorded by the scanner.

A new, more objective approach to computer training was developed for analysis of the three dates of data. Emphasis was placed on the standardization of a procedure for analysis of data. The procedure offered faster and consistently good duplication of attained results.

The results indicated an ability for automatic computer analysis-of remotely sensed multispectral scanner data to characterize and map land use categories within the test area. Additionally, results indicated an alteration of the data analysis procedure and land use classification scheme.

]]> Philip N. LeBlanc et al. Spectra of Normal and Nutrient-Deficient Maize Leaves http://docs.lib.purdue.edu/larstech/128 http://docs.lib.purdue.edu/larstech/128 Fri, 23 Oct 2009 15:56:12 PDT Reflectance, transmittance and absorptance spectral of "normal" and six types of nutrient-deficient (N, P, K, S, Mg, and Ca) maize (Zea mays L.) leaves were analyzed at 30 selected wavelengths from 500-2600 nm. The analysis of variance showed significant differences in reflectance, transmittance and absorptance in the visible wavelengths among leaf numbers 3, 4, and 5, among the seven treatments, and among the interactions of leaf number and treatments. In the infrared wavelengths only treatments produced significant differences.

The chlorophyll content of leaves was reduced in all nutrient deficient treatments. Percent moisture was increased in S-, Mg-, and N-deficiencies. Positive correlations were obtained (r = 0.7) between moisture content and percent absorption at both 1450 and 1930 nm. Polynomial regression analysis of leaf thickness and leaf moisture content showed that these two variables were significantly and directly related (R = 0.894). Leaves from the P- and Ca-deficient plants absorbed less energy in the near infrared than the normal plants; S-, Mg-, K-, and N-deficient leaves absorbed more than the normal.

Leaf thermograms were prepared on normal and S- and N-deficient leaves. Both S- and N-deficient leaves had higher temperatures than normal maize leaves.

]]> A. H. Al-Abbas et al. Determining Density of Maize Canopy: II. Airborne Multispectral Scanner Data http://docs.lib.purdue.edu/larstech/126 http://docs.lib.purdue.edu/larstech/126 Fri, 23 Oct 2009 15:56:11 PDT Multispectral scanner data were collected in two flights over a light colored soil background cover plot at an altitude of 305 m. Energy in eleven reflective wavelength bands from 0.45 to 2.6 µm was recorded by the scanner. Four growth stages of maize (Zea mays L.) were present at the time of each over flight, giving a wide range of canopy densities for each flight date. Leaf area index measurements were taken from twelve subplots at the time of each overflight, and were used as a measure of canopy density.

Ratio techniques were used to relate uncalibrated scanner response to leaf area index. The ratios of scanner data values for the 0.72 to 0.62 µm wavelength band over the 0.61 to 0.70 µm wavelength band were calculated for each plot. The ratios related very well to leaf area index for a given flight date but could not be generalized between data from different flights because of uncertainty in scanner response on different dates. The results indicated that spectral data from maize canopies could be of value in determining canopy density.

]]> E. R. Stoner et al. Determining Density of Maize Canopy: III. Temporal Considerations http://docs.lib.purdue.edu/larstech/127 http://docs.lib.purdue.edu/larstech/127 Fri, 23 Oct 2009 15:56:11 PDT Multispectral scanner data were collected In two flights over ground cover plots near the Purdue University Agronomy Farm's Weather Station at an altitude of 305 m. Energy in eleven reflective wavelength bands from 0.46 to 2.6 µm was recorded by the scanner. A set of eight ground reflectance panels was in close proximity to the ground cover plots and was used to normalize the scanner data obtained on different dates. The ground reflectance panels were used to relate laboratory reflectance measurements to scanner response. Separate prediction equations were obtained for both flight dates for all eleven reflective wavelength bands of the multispectral scanner. In this way, scanner response was normalized to ground panel reflectance. By normalizing the scanner data, ratios of scanner data could be related to leaf area index over time.

Normalized scanner data were used to plot relative reflectance versus wavelength for the ground cover plots. Spectral response curves resulted which were similar to those for bare soil and green vegetation as determined by laboratory measurements. The spectral response of different ground cover plots represented a "mixing" of the spectral response curves for the bare soil and green vegetation components of the scene. The spectral response curves from the normalized scanner data indicated that reflectance in the 0.72 to 1.3 µm wavelength range increased as leaf area index increased. A decrease in reflectance was observed in the 0.65 µm chlorophyll absorption band as leaf area index increased. This confined the validity of using the ratio of the response from a near infrared wavelength band to that of the red wavelength band in relating multispectral scanner data to leaf area index in maize.

]]> E. R. Stoner et al. Multispectral Determination of Vegetative Cover in Corn Crop Canopies http://docs.lib.purdue.edu/larstech/124 http://docs.lib.purdue.edu/larstech/124 Fri, 23 Oct 2009 15:56:10 PDT This research was designed to study the relationship between different amounts of vegetative ground cover and the energy reflected by corn canopies. Low altitude photography and an airborne multispectral scanner were used to measure this reflected energy.

]]> E. R. Stoner et al. Determining Density of Maize Canopy: I. Digitized Photography http://docs.lib.purdue.edu/larstech/125 http://docs.lib.purdue.edu/larstech/125 Fri, 23 Oct 2009 15:56:10 PDT This research was designed to study the relationship I between different densities of maize (Zea mays L.) canopies and the energy reflected by these canopies. Field plots were laid out, representing four growth stages of maize. Two plot locations were chosen, one on a dark soil and the other on a very light colored surface soil. Spectral and spatial data were obtained from color and color infrared photographs taken from a vertical distance of 10 m above the maize canopies. Estimates of ground cover were made from these photographs and were related to field measurements of leaf area index. Ground cover could be predicted from leaf area index measurements by a second order equation.

Color infrared photography proved helpful in determining the density of maize canopy on dark soils. Color photography was useful for determining canopy density on light colored soils, provided the percent ground cover did not exceed approximately 75%.

Microdensitometry and digitization of the three photographically separated dye layers of color infrared film showed that the near infrared dye layer is the most valuable (in canopy density determinations. Computer analysis of the digitized photography provided an accurate method of determining canopy density.

]]> E. R. Stoner et al. Mapping of Soils and Geologic Features with Data from Satellite-Borne Multispectral Scanners http://docs.lib.purdue.edu/larstech/122 http://docs.lib.purdue.edu/larstech/122 Fri, 23 Oct 2009 15:56:09 PDT The ERTS-1 satellite provides opportunity for quick inventory and assessment of geologic, soils, and vegetative cover aspects of large-scale areas of the Earth's surface. Collin County, Texas, U.S.A., a 2270 Km² area of relatively simple geology and soil associations was chosen for initial study, using ERTS-1 4-channel multispectral scanner data. These data were analyzed by computer-implemented pattern recognition techniques developed at the Purdue University Laboratory for Applications of Remote Sensing (LARS). The results indicate excellent visual correlation, on a gross scale, between automatically produced maps and existing geologic and soils maps and field information.

]]> M. F. Baumgardner et al. Application of Multispectral Remote Sensing to Soil Survey Research in Indiana http://docs.lib.purdue.edu/larstech/123 http://docs.lib.purdue.edu/larstech/123 Fri, 23 Oct 2009 15:56:09 PDT Recent advances have been made in the technology of measuring radiance from the earth's surface using multiple-wavelength airborne scanning spectrometers. Concurrently, advances were being made in the application of computer-implemented pattern recognition techniques to these multispectral data. Together these two tools have resulted in a capability for napping various earth surface features with extreme rapidity and varying degrees of accuracy. This study compared computer-implemented mappings based on spectral properties of bare soil surfaces with mapping units of interest to soil surveyors. Some soil types could be differentiated by their spectral properties. In other cases, soils with similar surface colors and textures could not be distinguished spectrally. The spectral maps seemed useful for delineating boundaries between soils in many cases.

]]> A. L. Zachary et al.