Showing papers on "Feature extraction published in 1980"

Decorrelation Methods of Texture Feature Extraction

Abstract: This paper presents the development and evaluation of a visual texture feature extraction method based on a stochastic field model of texture. Results of recent visual texture discrimination experiments are reviewed in order to establish necessary and sufficient conditions for texture features that are in agreement with human discrimination. A texture feature extraction technique involving autocorrelation function measurement of a texture field, combined with histogram representation of a statistically decorrelated version of the texture field, is introduced. The texture feature extraction method is evaluated in terms of a Bhattacharyya distance measure.

A Unifying View of Second-Order Measures of Quality for Signal Classification

Abstract: The major purpose of this paper is to promote interchange between the fields of pattern recognition and communications, in the realm of statistical classification. The general class of second-order measures of quality for statistical classification is defined. The variety of members in this class that have been used by practitioners or proposed by theorists for numerical pattern-classification and signal waveform-classification are compared and contrasted. The several measures that are the most generally applicable are shown to be either equivalent to each other or characterizable in terms of each other, thereby revealing an inherent unity. For example, the ratio of between-class-scatter to within-class-scatter used in pattern recognition and the ratio of signal-energy to noise-energy used in communications are unified through an identification of signal with between-class-scatter and noise with within-class-scatter. Results on equivalences are stated and proved for waveform classification rather than numerical classification in order to complement the extensive literature on the latter, and to emphasize applicability to communications. This entails introduction of a scatter ratio for waveforms. In a companion paper, second-order measures of quality are used as a basis for a general nearestprototype signal-classification methodology; canonical signal features for this methodology are identified, and a general approach for determining appropriate class prototypes is given. These two papers provide an integrated approach to the design of a complete signal classifier, i.e., feature extraction and discriminant-functional design tailored to fit a minimumdistance discrimination rule.

FAZYTAN: a system for fast automated cell segmentation, cell image analysis and feature extraction based on TV-image pickup and parallel processing.

Abstract: Cell location, segmentation and feature extraction of cell images are principal tasks of a high-resolution system for automated cytology. To perform these tasks with high speed, image processing algorithms and the architecture of a processor have to be optimized mutually. This has led to the development of a fast system for the evaluation of cytologic samples based on an optimized TV microscope, a host minicomputer with different peripheral array processors and digital image storages. The processors are optimized in speed for two-dimensional local operations to investigate neighborhood relations and morphology in cell images. Two-dimensional transformations of TV images (288 x 512 x 8 bit) can be carried out within 20 to 200 msec. The processors are able to realize linear filter functions (correlation, convolution) as well as nonlinear functions (median filtering). A set of measurements like area, circumference and connectivity can be derived parallely from one image in 20 msec. The system performs efficient and fast detection and segmentation of cells scanned in one TV frame within one second as well as the extraction of a large number of morphologic features within a few seconds. Based on these procedures, high-resolution analysis of several thousand cells of a sample within one minute will be possible.

Topographic feature extraction using sensor array system

Abstract: A method and apparatus for topographic feature extraction by masking a video signal representation of an optical image using a two-dimension Walsh function is disclosed. The optical energy of an image is converted into a video signal. A Walsh function signal in accordance with a two-dimension Walsh is generated. The video signal is multiplied with the Walsh function signal to produce a Walsh transform coefficients signal indicative of the decomposed spectral components of the image. The method and apparatus can further include providing selected topographical feature signals in accordance with the Walsh transform coefficients of selected topographic features. A selected topographical feature can be detected by comparing the Walsh transform coefficients signal with the selected topographical feature signals. The Walsh transform coefficients signal is unique to the selected topographical feature being detected.

Real-Time Medical Image Processing

Abstract: General.- Towards an Image Analysis Center for Medicine.- 1. Introduction.- 2. The Interactive Image Analysis System.- 2.1 The Input Units.- 2.2 The Image Display System.- 2.3 The Computer System.- 3. The Computerized Microscope.- 3.1 The Input System.- 3.2 The Computer System.- 3.3 The Microscope Stage Controller.- 4. Discussion.- 5. Acknowledgements.- 6. References.- Cellular Computers and Biomedical Image Processing.- 1. Introduction.- 2. Cellular Computers.- 2.1 The von Neumann Cellular Automaton.- 2.2 Cellular Automata and Image Processing.- 2.3 Pipeline Image Processor.- 3. Cellular Computers and Image Processing Research.- 3.1 Programming the Cellular Computer.- 3.2 Research Status.- 4. Biomedical Applications of Cellular Digital Image Processing.- 4.1 Coronary Artery Disease.- 4.2 Tissue Growth and Development.- 4.3 Genetic Mutagen Studies.- 5. References.- Radiology.- Ultra High-Speed Reconstruction Processors for X-Ray Computed Tomography of the Heart and Circulation.- 1. Introduction.- 2. Computation and Display.- 2.1 The Cone Beam Geometry.- 2.2 Choice of a Suitable Reconstruction Algorithm.- 2.3 Filtration Methods.- 2.4 Reconstruction Processor Design.- 2.5 Back-Projection Implementations.- 3. Flexibility of the High Speed Parallel Processing Reconstruction Architecture.- 4. Acknowledgements.- 5. References.- Computer Analysis of the Ultrasonic Echocardiogram.- 1. Introduction.- 2. System Configuration.- 3. Processing of Ultrasonic Echocardiogram.- 4. Clinical Results.- 5. Conclusions.- 6. Acknowledgements.- 7. References.- Toward Computed Detection of Nodules in Chest Radiographs.- 1. Introduction.- 2. Materials and Methods.- 2.1 Preprocessor.- 2.2 Circularity Detector.- 2.3 Boundary Follower.- 2.4 Classifier.- 3. Experimental Results.- 4. Conclusions.- 5. Acknowledgements.- 6. References.- A Parallel Processing System for the Three-Dimensional Display of Serial Tomograms.- 1. Introduction.- 2. System Outline.- 2.1 Tomograms.- 2.2 Functions.- 3. System Architecture.- 4. Image Input Device.- 5. Image Processing.- 5.1 Reduction of Size.- 5.2 Smoothing.- 5.3 Binarization.- 5.4 Differentiation.- 5.5 Rotation.- 6. Three-Dimensional.- 7. Illustrative Experiment.- 8. Concluding Remarks.- 9. Acknowledgements.- 10. References.- Dynamic Computed Tomography for the Heart.- 1. Introduction.- 2. Dynamic Scanner.- 2.1 CT Images Using the Dynamic Scanner.- 2.2 ECG Gated Image Using the Dynamic Scanner.- 2.3 ECG Phase Differentiation Method.- 2.4 Comparison of Methods.- 3. Proposed Method for Faster Synchronization.- 3.1 Multilens Method.- 3.2 Rocking Fan Beam Method.- 4. Conclusion.- 5. References.- Efficient Analysis of Dynamic Images Using Plans.- 1. Introduction.- 2. Hardware System.- 3. Analysis of Heart Wall Motion in Cine-Angiograms.- 3.1 Plan-Guided Analysis of Thickness of the Heart Wall.- 3.2 Input Pictures.- 3.3 Planning for Feature Extraction.- 3.4 Efficient Heuristic Search for Smooth Boundaries.- 3.5 Selection of Frames for Analysis.- 3.6 Analysis of Consecutive Frames.- 3.7 Measurement of Wall Thickness.- 4. Conclusion.- 5. Acknowledgements.- 6. References.- Real-Time Image Processing in CT-Convolver and Back Projector.- 1. Introduction.- 2. Image Processing in CT.- 3. System Configuration.- 4. High Speed Processor.- 4.1 Convolution.- 4.2 Back Projection.- 4.3 Processing Time.- 5. Display System.- 6. References.- Histology and Cytology.- Detection of the Spherical Size Distribution of Hormone Secretory Granules from Electron Micrographs.- 1. Introduction.- 2. Materials and Methods.- 2.1 Manual Method for Size Distribution Analysis.- 2.2 Computer Method for Size Distribution Analysis.- 2.3 Analysis.- 3. Results.- 4. Discussion.- 5. Conclusion.- 6. References.- The Abbott Laboratories ADC-500T.M..- 1. Introduction.- 1.1 Rationale.- 1.2 Rate-Limiting Factors.- 2. Sample Preparation.- 2.1 The Spinner.- 2.2 The Stainer.- 3. Real-Time Blood Cell Image Analysis.- 3.1 The Computer-Controlled Microscope.- 3.2 Cell Acquisition.- 3.3 High Resolution Data Analysis.- 3.4 Cell Classification.- 3.5 System Timing.- 3.6 Review.- 4. Results.- 5. Summary and Conclusion.- 6. Future Trends.- 7. References.- An Approach to Automated Cytotoxicity Testing by Means of Digital Image Processing.- 1. Introduction.- 2. The Principle of the Lymphocytotoxicity Test.- 3. Description of Our Instrument.- 3.1 Image Input Device.- 3.2 Autofocusing Algorithm.- 3.3 Thresholding.- 3.4 Binary Pattern Matching.- 3.5 Cell Counting.- 4. Important Parameters to Determine Positivity.- 4.1 Cell Size.- 4.2 Density.- 4.3 Halo.- 5. Experimental Results.- 6. Summary.- 7. Acknowledgements.- 8. References.- The diff3T.M. Analyzer: A Parallel/Serial Golay Image Processor.- 1. Introduction.- 2. Functional Organization.- 3. Automated Microscope.- 3.1 Optics.- 3.2 Optical Bench.- 4. Golay Image Processor.- 4.1 Golay Logic Processor (GLOPR).- 4.2 Application of GLOPR.- 5. Summary.- 6. References.- Computer-Aided Tissue Stereology.- 1. Introduction.- 2. Analysis of Muscle Biopsies.- 3. Analysis of the Placenta.- 4. Analysis of Adipose Tissue.- 5. Computer-Aided Data Capture.- 6. Choice of Method.- 7. Acknowledgements.- 8. References.- Interactive System for Medical Image Processing.- 1. Introduction.- 2. Using the SUPRPIC Interactive Image Processing System.- 2.1 Cellular Logic for Image Processing.- 2.2 The Command Structure of SUPRPIC.- 3. Examples of Using SUPRPIC.- 3.1 Digitizing the Image.- 3.2 Determination of Tissue Image Architecture.- 4. Summary and Conclusions.- 5. References.- Real-Time Image Processing in Automated Cytology.- 1. Introduction.- 2. System Design and Specifications.- 2.1 System Design.- 2.2 System Specifications.- 3. Image Input.- 3.1 Image Input System Design.- 3.2 Image Sensor Module.- 3.3 Screening Module.- 3.4 Automated Focusing Module.- 4. Feature Extraction Method.- 4.1 Feature Extraction Hardware.- 4.2 Feature Extraction Procedure.- 5. Processing Sequence.- 6. Conclusion.- 7. References.- The Development of a New Model Cyto-Prescreener for Cervical Cancer.- 1. Introduction.- 2. Fundamental Ideas.- 2.1 Improvement of Diagnostic Performance.- 2.2 Improvement of Processing Speed.- 2.3 Automatic Slide Preparation.- 3. Improvement of Diagnostic Performance.- 3.1 Improvement of Cell Image Quality.- 3.2 Addition of Morphological Feature Parameters.- 3.3 Improvement of Preprocessing Program for Feature Extraction.- 3.4 Increase in Number of Cells Analyzed.- 4. Hardware Implementation.- 4.1 High Resolution TV Camera.- 4.2 Automatic Focusing Mechanism.- 4.3 High Speed Image Processor.- 4.4 Flexible Controllers.- 4.5 Automatic Smear Preparation.- 5. Conclusion.- 6. Acknowledgement.- 7. References.- Author Index.

Image database management for pattern information processing studies

Abstract: The image database EIDES was developed at the Electrotechnical Laboratory. It contains a considerable number of standard images for experimental studies on pattern recognition and image processing. The file structure of EIDES is based on the ‘Standard Format for Digital Images in Japan.’ This paper describes the management software of EIDES. Many additional subroutines exist for format conversion or access to image data in user's image processing programs. Especially, a set of executive routines can be used for manipulating large images stored in EIDES. These routines free each primitive image processing routine designed for incore use from data management work. This assures transportability between systems.

Comparative study of optical-digital vs all-digital techniques in textural pattern recognition

Abstract: In this study we examine the application of both optical–digital and all-digital techniques in textural pattern recognition and make a comparison of the two approaches. The optical–digital scheme makes use of an optical–digital computer to generate textural measurements based on the 2-D irradiance spectrum. The all-digital scheme produces measurements based on gray-tone spatial-dependence matrices. In both cases two feature extraction algorithms were employed: the Hotelling trace method and the Foley-Sammon discriminant vector analysis. Classification was accomplished using the k-nearest neighbor decision rule. The performance of these techniques was evaluated in an experiment involving the classification of four texture patterns. The results show that, for the textures chosen, both approaches give high classification accuracy with the optical–digital method performing somewhat better.

Dedicated LSI for a Microprocessor-Controlled Hand-Carried OCR System

Abstract: The binary picture processing and recognizing stages of an optical character recognition (OCR) system have been designed using both flexibility of available microprocessors and speed of peripheral custom-designed integrated circuits, A dedicated Iarge-scale integrated (LSI) processor performs edge detection and thinning of a 32 X 24 digitied one-piece pattern. The output signal-a set of 3 bit vectors describing the skeletonized character contour-feeds a microprocessor which controls the character recognition algorithm including pattern segmentation, filtering, feature extraction, and classification decision. This low-cost equipment is especirdly suitable for hand-carried OCR systems where well-formed printed alphanumerics are to be read: However, continously deformed patterns like carefully handprinted characters are recognized as well. A system reading speed of 100 characters/s (or 30 cm/s) can be achieved.

Image spectrum analyzer for cartographic feature extraction

Abstract: An image spectrum analyzer and method for cartographic feature extraction comprising an image, means for generating a Walsh function light pattern and for illuminating the image with the Walsh function light pattern to procuce a masked pattern, and means for receiving the masked pattern and for producing an electrical signal in accordance with same. A digital signal can be provided in accordance with the electrical signal, and a digital data processor under stored program control can provide a Walsh transform coefficient representative of the digital signal. More than one Walsh function light pattern can be generated by the apparatus and method of the present invention to produce more than one masked pattern, and representative Walsh coefficient. The Walsh coefficients can be used to perform cartographic feature extraction of the image.

Speech Signal Processing and Feature Extraction

Abstract: Speech signal processing and feature extraction is the initial stage of any speech recognition system; it is through this component that the system views the speech signal itself This chapter introduces general approaches to signal processing and feature extraction and surveys the techniques currently available in these areas

Automatic Visual Inspection Of Reed Switches

Abstract: In many production chains visual inspection of products is an important manufacturing consideration with respect to quality control. Recent progress in image processing and pattern recognition led the way to economically justified applications; modern technology enables the construction of such auto mata featuring high reliability and constancy. This paper describes an automatic visual inspection machine performing intelligent control tasks in a very short time. The possible applications include dimension control of products; inspection of objects on shape, greyness, or texture; sorting of objects; positioning; etc. The design is based on a fast synthesis procedure of a TV image by profile extraction. The features to be controlled are extracted from these profiles and compared with upper/lower limits obtained from a learning process. The system performances are expressed by the elaboration of an industrial case, the real-time visual inspection of reed switches. More than 35 features are detected and controlled for each switch within 1 second. Several experiments with a prototype version have proven the feasible use as an industrial on-line controller.

Performance of the Discrete Fourier Transform Satellite Imagery Classification Technique

Abstract: : The discrete Fourier transform (DFT) automated satellite imagery classification technique is designed to detect and identify cloud features from 25 x 25 nautical mile (nm) Defense Meteorological Satellite Program (DMSP) visible and infrared imagery samples. The DFT classifier performs two basic steps: feature extraction and sample identification. The feature extraction technique reduces and compresses the original visible and infrared data of an imagery sample in order to retain only the information needed to determine the cloud type of that sample. A multivariate normal density discriminant function then uses this information to identify the cloud type of the image sample from nine possible classes. The results obtained through this experiment show that the classifier can provide accurate and reliable cloud type classifications of satellite imagery samples. This report presents a detailed mathematical description of the DFT spectral classifier technique and offers some ideas for future modifications to the classifier. (Author)

Feature extraction and recognition of Japanese voiceless stop consonants by principal component analysis

Abstract: At the present time, how to extract acoustic features of voiceless stop consonants is oneof the most difficult problems remaining unsolved in the field of automatic speech recognition.This paper describes a study on characteristic features of the Japanese voicelessstop consonants /p/, /t/ and /k/. A multi-dimensional statistical analysis method isapplied to analyze their spectra. Analysis reveals that the principal characteristics discriminatingbetween them are reflected in the accumulated power from about 1kHz upto 5kHz and the existence of a spectral peak. It is also found that these feature parametersmake it possible to separate the voiceless stop consonants from each other. Experimentsof automatic recognition based on the maximum likelihood method are alsoperformed. They are carried out on condition that the following vowel is correctlyknown beforehand and that there are no errors in detection of the noise onset. It isfound that a recognition rate as high as about 97% can be attained for the training dataset of known utterances if the recognition algorithm designed to make the best use ofthe transition patterns of the feature parameters is adopted. It is also shown that theaccurate detection of the moment of noise burst is essential to attain a high recognition rate.

Feature Extraction From Forward Looking Infrared (FLIR) Imagery

Abstract: Many image processing tasks require the ability to extract useful features from digital images. The selection of a set of appropriate attributes to be extracted from an image constitutes a major problem. This paper examines a number of region features for their utility in target detection and object recognition. A simple hierarchical detection scheme is constructed which can be easily extended to perform classification. Experimental results are presented for FLIR images of tactical targets.© (1980) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Estimation of the performance of an array-processor oriented system for automatic Pap smear analysis.

Abstract: The rapid progress in computer technology makes possible the automatic analysis of thousands of cells on a slide in the field of automatic uterine cancer cytology. Our approach starts with the high-resolution scanning of visually selected and classified single cells determining the training set for discriminant analysis. On the basis of more than 15 morphologic and textural features measured, correct classification results of 95% are reached. Our aim is to speed up our image processing system by means of an Array-Processor, the model AP 120 B from Floating Point Systems. In this study a realistic time estimation of the scanning, segmentation, feature extraction and classification of about 40,000 cells on a slide is performed.

Statistical filter for image feature extraction.

Abstract: Edge extraction techniques have become important as a preprocessing step in extraction of image features for the purpose of image segmentation, object identification, and bandwidth compression. The use of conventional edge extractors such as Sobel and Laplacian filters results in images that in many cases have a high degree of clutter due to the natural spatial texture of the scene background. To overcome this difficulty, a statistical filter has been developed that enhances local grey level activity around objects while reducing contributions due to background. The statistical filter is employed in a neighborhood modification process where the central pixel is replaced with the third central moment computed from the surrounding neighborhood. Choice of the third central moment is due in part to the fact that it is a function of the scene within the neighborhood rather than the power spectral density (Wiener spectrum) of the neighborhood. Application of the filter requires no prior knowledge, and pixels within the filter window may be chosen in random order due to the statistical nature of the operation. Results of the filter applied to IR images show performance comparable with, and in some cases superior to, the Sobel and Laplacian filters most commonly used for feature and edge extraction.

Feature extracing system of binary pattern

Abstract: PURPOSE: To extract various high-degree feature quantity by extracting the various types of the respective features of a binary pattern, cutting out the regions of these features and coding these, obtaining the feature quantity of each region based on these coded regions. CONSTITUTION: A block of black is cut-out by a run length counter with respect to a binary coded input line pattern, and the beginning point coordinates thereof and the end point coordinates are made into a graph. Also, the local features of the continuous two lines or two rows of the binary pattern are extracted. For this, the noise processing in the line direction is performed and simultaneously, the correction of the local feature graph is carried out according to the noise processing. Next, the integrating process differing with each feature extracted is carried out, whereby large area feature extraction is carried out. As a result, the feature extraction and cut-out of the region are accomplished simultaneously and the results thereof are coded by run length. At the same instant of the large area feature extraction, the processing of the noise in the row direction is carried out. Next, the feature quantity of each region is extracted by the coded information. Thereby, the region is correctly cut out and the information remains without any loss; hence the extraction of various high-degree features is made feasible. COPYRIGHT: (C)1981,JPO&Japio

Dedicated LSI for a Microprocessor-Controlled Hand-Carried OCR System

Abstract: The binary picture processing and recognizing stages of an optical character recognition (OCR) system have been designed using both flexibility of available microprocessors and speed of peripheral custom-designed integrated circuits. A dedicated large-scale integrated (LSI) processor performs edge detection and thinning of a 32 × 24 digitized one-piece pattern. The output signal–a set of 3 bit vectors describing the skeletonized character contour–feeds a microprocessor which controls the character recognition algorithm including pattern segmentation, filtering, feature extraction, and classification decision. This low-cost equipment is especiaUy suitable for hand-carried OCR systems where well-formed printed alphanumerics are to be read. However, continously deformed patterns like carefully handprinted characters are recognized as well. A system reading speed of 100 characters/s (or 30 cm/s) can be achieved.

Character recognition system

Abstract: PURPOSE:To recognize characters easily by agreement between a feature extraction bit pattrn and memory contents of the operator's dictionary, by loading previously a personal dictionary into an ORC unit to generate a dictionary and by reading a form corresponding to the operator and by sending the feature extraction bit pattern to the dictionary and by collating the feature extraction bit pattern with memory contents of the operator's dictionary. CONSTITUTION:Character samples A11, B12, and C13 are inserted to read mechanism 14 of OCR unit 20 individually, and features are extracted by feature extraction part 15, and feature groups are stored in a memory as bit patterns to generate dictionary 16. Dictionaries A17, B18 and C19 for respective persons are generated from bit patterns for respective persons and are stored in floppy disc device 24. Here, in case that respective form entering persons A, B and C process forms A21, B22 and C23 entered by themselves, dictionaries of respective persons are loaded previously into the memory in OCR unit 30 to generate dictionary 27, and forms 21...23 are read, and features are extracted in 26 to send bit patterns to dictionary 27, and the bit pattern is collated with the bit pattern group from unit 24, and thus, character can be recognized easily due to coincidence between them.

Real-time image processing in automated cytology

Abstract: Uterine cancer is detected in its early stages by screening cells scraped from the uterus and spread on a microscope slide for visual observation. Cells are roughly screened by a cytological technician called a “screener.” A physician then makes the final diagnosis. Cyto-screening has been automated by us using image processing and pattern recognition techniques. The authors have made significant improvements in the following factors: (1) screening accuracy, (2) screening speed, and (3) sample preparation. New cytologic feature parameters and a hierarchical classification algorithm have been developed. As feature parameters we use nuclear area, nuclear area/cytoplasm area, and nuclear area with higher density. The hierarchical classification algorithm incorporates the cell growth method of Suzuki (1978). Screening speed is expected to be twice that achieved in manual screening. The authors aimed at a speed of 5 min./sample and have succeeded in developing a system with a speed of 2 min./sample.

Some Basic Techniques in Picture Processing and Feature Extraction

Abstract: Image understanding systems for the analysis of complex scenes should have highly developed abilities in picture processing and feature extraction to process digital images and measure the various properties of regions and lines. Especially in the analysis of aerial photographs, we have to calculate many different features to characterize a variety of objects: boundary smoothness for crop fields, elongatedness for roads, rivers and railroads, squareness for houses and buildings, textural properties for grasslands and forest areas, etc. The quality of these measurements, obtained by picture processing and feature extraction, has crucial effects on the higher-level recognition process. Thus, the early stages of feature extraction must be as accurate as possible to obtain good final results.

Character decision system of character recognition unit

Abstract: PURPOSE:To alternate between reduction in the number of unreadable characters and extreme reduction in error read by providing the recognizing decision part of a character recognition part with primary and secondary decision parts and by sending the output of either one by an external command CONSTITUTION:A character pattern from character pattern register 11 is scanned by feature extraction part 12 to extract various feature values (fj) (j=1-m), which are inputted to primary decision part 13 and secondary decision part 14 When a feature value satisfying category (ci) among feature values fj is detected, decision part 13 outputs primary recognition result (cfi) Decision part 14 extracts a feature value required to reduce extremely the error read of the category decided by decision part 13 and output (csi) With external terminal T' OFF, AND circuit 10i and OR circuit 30i output signal (cfi) and with terminal T' ON, output result (csi) of decision part 14 is outputted simultaneously; at this time, signal (csi) becomes output (ci) via AND circuit 20i and OR circuit 30i

The Engineer Topographic Laboratories (ETL) Hybrid Optical/Digital Image Processor

Abstract: An optical-digital processor for generalized image enhancement and filtering has been designed and is now under construction. The optical subsystem is a two PROM Fourier filter processor. Input imagery is isolated, scaled, and imaged onto the first PROM. This input plane acts like a liquid gate and serves as an incoherent to coherent converter. The image is transformed onto a second PROM which also serves as a filter medium. Filters are written onto the second PROM with a laser scan-ner in real time. A solid state CCTV camera records the filtered image which is then digitized and stored in a digital image processor. The operator can then manipulate the filtered image using the gray scale and color remapping capabilities of the video processor as well as the digital processing capabilities of the mini-computer. The operator can then try new optical filters and iteratively develop optimum methods of detecting patterns. The goal of this research is to develop automated feature extraction algorithms which will minimize the need for human intervention. This system is currently being assembled at ETL, Fort Belvoir, VA.

Inventory estimation on the massively parallel processor

Abstract: This paper describes algorithms for efficiently computing inventory estimates from satellite based images. The algorithms incorporate a one dimensional feature extraction which optimizes the pairwise sum of Fisher distances. Biases are eliminated with a premultiplication by the inverse of the analytically derived error matrix. The technique is demonstrated with a numerical example using statistics obtained from an actual Landsat scene. Attention was given to implementation of the Massively Parallel processor (MPP). A timing analysis demonstrates that the inventory estimation can be performed an order of magnitude faster on the MPP than on a conventional serial machine.