GRIPS 2: Government Raster Image Processing Software & Data

home *** CD-ROM | disk | FTP | other *** search

/ GRIPS 2: Government Rast…rocessing Software & Data / GRIPS_2.cdr / dos / pc_mips / pc_mips.rmf < prev next >

Wrap

Text File | 1990-06-25 | 55.4 KB | 1,585 lines

Draft version of PC-MIPS User Manual as of 04/19/90 PC-MIPS User Manual Table of Contents PC-MIPS Overview . . . . . . . . . . . . . . . . . . . 1 Installing PC-MIPS . . . . . . . . . . . . . . . . . . 5 User Interface . . . . . . . . . . . . . . . . . . . . 5 Image Data Structure . . . . . . . . . . . . . . . . . . 6 User Input . . . . . . . . . . . . . . . . . . . . . . 7 PC-MIPS Programs CMBIMG . . . . . . . . . . . . . . . . . . . . . . . . 8 DERIV . . . . . . . . . . . . . . . . . . . . . . . . 8 EDGENH . . . . . . . . . . . . . . . . . . . . . . . . 9 EXT_ROI . . . . . . . . . . . . . . . . . . . . . . . . 10 FLT8B . . . . . . . . . . . . . . . . . . . . . . . . * IMDISP . . . . . . . . . . . . . . . . . . . . . . . . 10 LIST . . . . . . . . . . . . . . . . . . . . . . . . 10 MAPIMG . . . . . . . . . . . . . . . . . . . . . . . . 11 MAPLN1 . . . . . . . . . . . . . . . . . . . . . . . . 12 MDGRAD . . . . . . . . . . . . . . . . . . . . . . . . 13 MEDIAN . . . . . . . . . . . . . . . . . . . . . . . . 13 MIXBDS . . . . . . . . . . . . . . . . . . . . . . . . * NUVU . . . . . . . . . . . . . . . . . . . . . . . . * OVRLAY . . . . . . . . . . . . . . . . . . . . . . . . 14 PCA . . . . . . . . . . . . . . . . . . . . . . . . * PICSTR . . . . . . . . . . . . . . . . . . . . . . . . * RADARS . . . . . . . . . . . . . . . . . . . . . . . . * RATIO . . . . . . . . . . . . . . . . . . . . . . . . 16 ROI . . . . . . . . . . . . . . . . . . . . . . . . 17 ROTATE . . . . . . . . . . . . . . . . . . . . . . . . 17 SALTP . . . . . . . . . . . . . . . . . . . . . . . . 18 SCALE . . . . . . . . . . . . . . . . . . . . . . . . * SHAD2 . . . . . . . . . . . . . . . . . . . . . . . . 19 SHADRF . . . . . . . . . . . . . . . . . . . . . . . . * STATS . . . . . . . . . . . . . . . . . . . . . . . . * STEEP . . . . . . . . . . . . . . . . . . . . . . . . * STENCL . . . . . . . . . . . . . . . . . . . . . . . . * STRTCH . . . . . . . . . . . . . . . . . . . . . . . . 20 SUBIMG . . . . . . . . . . . . . . . . . . . . . . . . 21 SUN . . . . . . . . . . . . . . . . . . . . . . . . * WTCMBO . . . . . . . . . . . . . . . . . . . . . . . . 21 * - program write-up, as yet, not completed. PC-MIPS User Manual PC-MIPS Overview Historically, the processing and analysis of GLORIA side- scan sonar data within the U.S. Geological Survey has been done on DEC mini-computers using the MIPS software system. Due to the volume of work generated by the EEZ sea-floor mapping program, these systems have not been readily available to a large group of marine scientists who would like to perform investigations that require image analysis capabilities. At the same time, the availability and use of very powerful PCs has significantly increased. Recently, the processed GLORIA imagery has been made available on Compact Disks (CDROM) which can be accessed by PCs. These factors make a PC-based MIPS-like system highly desirable. The PC-MIPS system is an attempt to satisfy this need. PC-MIPS is an integrated collection of programs which have been adapted to a PC environment. Most of these programs have been selected from the MIPS software system developed at the USGS Image Processing Facility in Flagstaff, Arizona. While internal program and data structures have been extensively modified, an attempt has been made to preserve the basic appearance to the experienced MIPS user. PC-MIPS was developed using MS-DOS version 3.31 for the Compaq Deskpro. Microsoft's Fortran version 5.0 was chosen as the development language since about half of the MIPS system was written in Fortran (the other half being in the MAC assembler language). The selection of a display mechanism has largely been based on the availability of existing PC-based software. During the early stages of the GLORIA CD-ROM project, an image display program (IMDISP) developed by NASA at the Jet Propulsion Lab was evaluated and found to be excellent for displaying GLORIA images on a PC. IMDISP supports all the major graphics standards (i.e., CGA, EGA, VGA, and PGA). It provides an extensive set of commands and utilizes a very simple but robust set of file structures. The hardware utilized for the development of PC-MIPS is a Compaq 386/20 Model 130 equipped with 2Mb ram, the 80387/20 math coprocessor and a VGA color monitor. This configuration provides processing speeds which average 26Kb per second. In other words, if you process an input image 1Mb in size and create a resultant output image of the same size, the processing time will usually be under 1.5 minutes. Standard VGA provides a display of 16 colors and/or grey levels with a resolution of 640 by 480 pixels. A display of 64 grey levels can be achieved using the Orchid ProDesigner PLUS VGA card with 512K on-board memory. The hardware configuration which was used for development is recommended for use; however, PC-MIPS will run under MS-DOS on 1 PC-MIPS User Manual any generic PC/AT class machine with VGA graphics, and a hard disk. The software itself occupies approximately 2.6Mb. Realistically, a hard disk of at least 40Mb should be used. Following is a list of the programs currently available: CMBIMG Combines from 2 - 4 input images to create a composite output image which can be displayed via IMDISP. DERIV Generates the first differences in the horizontal, vertical, or diagonal direction. EDGENH Uses a Laplacian transform to provide edge enhancement capabilities. EXT_ROI Used to extract an irregularly shaped sub-image (defined in POLYGON.DAT) from up to 7 input images. FLT8B Boxcar spatial filtering (powerful workhorse; high and low pass filtering; surface generation; interpolation for gaps; edge enhancement; etc.). IMDISP JPL's image display program. Currently 18 commands are implemented. Among them are Profile, Stretch, Palette, Histogram, Browse, and Display with or without sub-sampling and/or zoom. Text may, also, be positioned on the display. LIST Generates a printout (LIST.PRT) of DN values for an image. Sub-image areas may be specified. MAPIMG Performs a user-specified mapping of 1 to 7 input files. An extremely robust program that generates an output image which is the result of a user-specified functional mapping applied to the input files. MAPLN1 Enhances linear features using a semi-linear line detecting/mapping technique (useful for regular image data as well as Topo, Gravity, etc.). MDGRAD Modified gradient method that instead of multiplying the constants together adds them because of the possibility of one of them being ZERO (linear feature enhancement/mapping technique; used on image and surface data such as TOPO). MEDIAN Applies a 3 by 3 median (middle value) filter. Will help remove high frequency noise. MEDIAN is particularly suitable for data that has a skewed distribution. However, the value obtained for the median may not be representative if the individual items do not tend to cluster at the center of the distribution. 2 PC-MIPS User Manual MIXBDS Will generate three output bands from the TM six bands by selecting the user-specified three optimum bands for each of vegetation, soils, and water. For example, vegetation pixels might have TM bands 2,4, and 5; soils pixels might have TM bands 1,4, and 5; while water pixels might have TM bands 1,2, and 3. This will allow the three output bands to be used for either color compositing or digital processing so that these three 'new' bands have the best combination for these three major cover types. NUVU Creates oblique views of image data using a Topographic height file and/or image file. If only the image file is used, the DN value is interpreted as height. OVRLAY Takes from 2 to 6 input files/images and generates one output file/image that indicates which of the input files have data in user specified ranges for the input files being used. PCA Generates the linear combinations for Principal Component Analysis of two to seven files (the output of STATS may be used as input coefficients). PICSTR This program allows the user to let the machine pick good stretch pairs for an image based on either min/ max parameters or by giving a min/max percent of the image. Values outside these ranges are ignored when trying to pick DN values. RADARS Uses digital elevation data as input (dma, dem, etc.) and outputs the elevation data with it's geometry converted to slant range mode. The user specifies the near range (nr) depression angle (used to determine nr to nadir distance) and the imaging system's height and pixel resolution. It assumes the radar scan direction is in the horizontal direction (i.e., the imaging system is on the left or right side of the image). RATIO Generates up to three output images that are the result of dividing the up to four input images by each other. ROI Used to define and extract an irregularly shaped sub- image from up to 7 input images. Requires VGA and serial MS compatible mouse. ROTATE Rotates an image 90 degrees in the clockwise or counter-clockwise direction. SALTP Does bit error or random noise removal (Salt and Pepper). 3 PC-MIPS User Manual SCALE Changes the scale of an image in the line direction, sample direction, or both. SHAD2 Does automatic shading correction on Sonar, Radar, or any other file. SHADRF Computes reflectance for any sun elevation and one of four sun directions using Lommel-Seeliger Law. STATS Computes the Mean, Standard Deviation, Variance, Covariance, and Correlation Coefficients between two to seven images. It will also create the eigen vector matrix that is required as input to PCA. STEEP Uses surface data as input (dem, dma, gravity, etc.) and generates a thematic map (0's or 255's) of ridges or drainage patterns (i.e., maps linear features). STENCL Uses two input files. The first file is 'stenciled' based on data of second file. Generates one output file. STRTCH Performs contrast stretching on an image using up to 10 user-specified stretch intervals. SUBIMG Allows the extraction and/or "flipping" (i.e., reversing) of sub-images in the line and/or sample direction. SUN Computes reflectance given any sun elevation and direction. WTCMBO Generates the weighted sum of 1 to 4 input files. The weighting coefficients are specified by the user and must be real numbers. The DN values for the output file are computed as follows: OUT = C1*INPUT1 + C2*INPUT2 + C3*INPUT3 + C4*INPUT4 + C5 Where C1, C2, C3, C4, and C5 are user-specified constants. 4 PC-MIPS User Manual Installing PC-MIPS PC-MIPS can be installed on any hard disk drive (e.g., C:, D:, E:,,,etc.). Once a drive is selected, the installer should create a root-level directory named \MIPS. Next, the two sub- directories \MIPS\BIN and \MIPS\AUX_FILE must be created. These directories are created using the DOS command MKDIR or MD. The executable programs of PC-MIPS are stored in the \MIPS\BIN directory. The \MIPS\AUX_FILE directory contains various auxiliary files used by PC-MIPS programs. Two 1.2Mb floppy disks contain three programs which are used to load the \MIPS sub-directories. These programs are named AUX_LOAD.EXE, BIN_LD1.EXE, and BIN_LD2.EXE (they were created using the program PKSFX from PKWARE INC., 7032 Ardara Ave., Glendale, WI 53209). The installer should first load the AUX_FILE sub-directory by typing A:AUX_LOAD \MIPS\AUX_FILE. Next, the BIN sub-directory is partially loaded by typing A:BIN_LD1 \MIPS\BIN. The installation is completed by inserting the second floppy disk and typing A:BIN_LD2 \MIPS\BIN. These steps can be performed automatically by executing the INSTALL.BAT files found on each floppy disk. They assume installation on hard disk drive C:. Once the two floppy disks have been loaded, the AUTOEXEC.BAT file should be changed so that the display program IMDISP knows what display device is being used. This is done by setting the environmental variable IMDISP. For instance, if the hardware has standard VGA capabilities, the line SET IMDISP=VGA must be added to the AUTOEXEC.BAT file. A complete list of devices supported by IMDISP can be found in the file IMDISP.DOC. PC-MIPS can now be executed by typing \MIPS\BIN\PC_MIPS. If \MIPS\BIN is added to the DOS PATH command in the AUTOEXEC.BAT file, the system can be run by simply typing PC_MIPS. User Interface Individual PC-MIPS programs can be executed directly or via a menu-driven interface. A menu-driven interface presents the user with a list of the PC-MIPS programs available for use and provides a mechanism for selecting and executing individual entries. Individual programs in the list may be selected by name or number. PC_MIPS, also, provides a gateway which allows execution of DOS commands without exiting the system. This gateway is entered by typing DOS at the user prompt. After the gateway is entered, the user returns to the menu by typing PC_MIPS in response to the DOS> prompt. 5 PC-MIPS User Manual Image Data Structure The PC-MIPS image data structure is a subset of JPL's PDS data structure (see IMDISP.DOC). This subset has the following restrictions: 1. Only 8 bit data is supported. 2. The image definition must be contained in a separate image label file. 3. Image header records are not supported. 4. Line prefix and suffix bytes are not supported. A PC-MIPS image is composed of two files. The first is an image label file and the second is an image data file. Image label files contain information describing the image such as the number of lines, the number of samples, and the location (path name) of the image data file. Records of the image label file contain keywords and values. The mandatory keywords are FILE_TYPE, IMAGE_LINES, LINE_SAMPLES, IMAGE_POINTER, and END. These are described as follows: FILE_TYPE - Must have the value IMAGE. IMAGE_LINES - Gives the number of lines (i.e., rows) in the image. LINE_SAMPLES - Gives the number of samples (i.e., columns) per line. IMAGE_POINTER - Gives the path name of the image data file. END - Must be the last record of a label file. If the PDS keyword LINE_PREFIX_BYTES is given, its value will be added to the LINE_SAMPLES value. The following image label file named SAMPLE defines an image of 460 lines and 407 samples: FILE_TYPE = IMAGE IMAGE_LINES = 460 LINE_SAMPLES = 407 IMAGE_POINTER= \images\sample.img END The image data file is binary (i.e., non-printable) and contains the DN values which comprise the image. This file MUST be IMAGE_LINES times LINE_SAMPLES bytes in size. It is simply a contiguous sequence of bytes (i.e., DN values) which represent the image stored in row (i.e., line) major order. While running PC-MIPS programs, the user refers to an image using the image label file name. THIS FILE NAME SHOULD NOT HAVE A DOS EXTENSION (i.e., not contain a period). A PC-MIPS output image will be composed of the user-specified image label file name and an image data file whose name is the same but has .IMG as a DOS extension. 6 PC-MIPS User Manual User Input PC-MIPS has two types of user input. The first type is file names (input and/or output). The second type is parameters that control the computation. File names are entered by the user at the FILES> prompt. The user responds by entering a list of output names, followed by an equal sign, followed by a list of input names. Individual elements (file names) in these lists are separated by commas. File names may not have extensions (i.e., may not contain a period) and full DOS path names may be used. For example, FILES>OUT1,OUT2=INPUT1,INPUT2,INPUT3 represents lists of two output files (OUT1 and OUT2) and three input files (INPUT1, INPUT2, and INPUT3). For programs that generate no output files, the user only enters the list of input files. If an input file does not exist, the user is informed and requested to reenter the FILES> line. If an output file already exists, the user is asked if this file should be over-written. Prior to the FILES> prompt the user is informed of the expected number of file names and the maximum number of samples per line that is permitted. All input files must be of the same size (i.e., contain the same number of lines and samples). The input of program parameters is given by the user at the PARAM> prompt. Prior to issuing the PARAM> prompt, the program displays the valid parameter names, their default values (enclosed in brackets), and their meaning. Parameters are assigned values by entering their name, followed by the equal sign, followed by a value. Values may be numeric constants or character strings. Character strings must be enclosed in apostrophes. Some parameters are set by simply entering their name (i.e., without the equal sign followed by a value). An example would be the specification of which type of computation is desired. Commas are used to separate the setting of multiple parameters. For example, the valid parameters for the LIST program are SS, SL, ES, EL, LINC, SINC, PAGE, and HEAD. If the user entered the line SS=200, EL=200, PAGE, HEAD='1st 200' he would be requesting a single page listing of the first 200 lines and samples with the title '1st 200'. 7 PC-MIPS User Manual PC-MIPS Programs CMBIMG This program is used to create an output file composed of multiple images which can be viewed using the IMDISP program. CMBIMG prompts for the input/output file specifications in normal PC-MIPS form (i.e., OUTPUT=INPUT1,INPUT2,INPUT3,INPUT4). There may be from two to four input images and the size (number of lines and samples) of all input images MUST be equal. The composite output image (i.e., file) is structured as shown below: ------ ------ |INPUT1|INPUT2| OUTPUT= |------|------| |INPUT3|INPUT4| ------ ------ hint: An output file of more than four input images can be created by repeated use of CMBIMG. Example: CMBIMG 18-MAY-88 # of files bit types: 8 2 - 4 input max samples: 3000 1 output File name where the processed results will be stored. (e.g., OUTPUT File name(s) that has (have) the input data to be processed. =INPUT1,INPUT2,INPUT3) ------ ------ |INPUT1|INPUT2| OUTPUT= |------|------| |INPUT3|INPUT4| ------ ------ FILES> comb=sample,m1,m2 FILES> Do you want to display comb (y/n)? n DERIV Program DERIV performs a derivative (or first difference) function on an image. The derivative can be in the horizontal, vertical, or diagonal direction as specified by the user. The equations used are: horizontal - Y(i,j) = X(i,j) - X(i+1,j) + C vertical - Y(i,j) = X(i,j) - X(i,j+1) + C diagonal - Y(i,j) = X(i,j) - X(i+1,j+1) + C 8 PC-MIPS User Manual where X is the input image, Y is the output image, and C is a user-specified constant. The user enters the input/output file names at the FILES> prompt. The program then displays the four valid parameters and issues a PARAM> prompt. The desired direction is selected entering one of the three directional parameters HORZ, VERT, or DIAG. The fourth parameter IADB is provided as an add-back constant which defaults to 127 if not specified. Example: DERIVE 12-JAN-85 # of files bit types: 8 1 input max samples: 3000 1 output Filename where the processed results will be stored. (e.g., OUTPUT Filename that has the input data to be processed. =INPUT) FILES> hdrev=sample Valid parameters are: HORZ:[] - Performs a horizontal derivative VERT:[] - Performs a vertical derivative DIAG:[] - Performs a diagonal derivative IADB:[127] - Addback constant (e.g., HORZ,IADB=127) PARAM> horz, iadb=100 FILES> Do you want to display hderv (y/n)? n EDGENH EDGENH uses a Laplace transform to perform edge enhancement on images. The transform matrix is as follows: 0 -1 0 -1 4 -1 0 -1 0 The user is prompted for input/output file names via the FILES> prompt. Next, the user is prompted to enter a value representing the fraction of the Laplace transform to be added. Each interior pixel Y(i,j) is calculated according to the formula Y(i,j)=X(i,j)+f*[4*X(i,j)-X(i-1,j)-X(i,j-1)-X(i+1,j)-X(i,j+1)] where X is the input image, Y is the output image and f is the user-specified fraction. Boundary pixels are unchanged. Example: EDGENH 18-JUNE-88 # of files bit types: 8 1 input 9 PC-MIPS User Manual max samples: 3000 1 output File name where the results are to be stored (e.g., OUTPUT File name where the input data is located = INPUT) FILES> edge = sample Enter fraction of Laplace transform to be added: 0.4 FILES> Do you want to display edge (y/n)? n EXT_ROI This program extracts sub-images from up to 7 input images. The sub-image is specified by an ascii file named POLYGON.DAT which contains the polygon vertices which define the sub-image. The first record in this file contains the number of vertices. Subsequent records contain the vertices themselves. They are given as sample/row pairs. This file must be present in the working directory when the program is executed. Output and input file names are given by the user at the FILES> prompt. IMDISP Refer to IMDISP.DOC file for description. LIST This program will generate a printout of DN values. The user can select a window by specifying starting/ending line and sample numbers. Additionally, line and sample increments may be given. If a single-page printout is requested, the program automatically determines the line and sample increments which will achieve this result. The name of the image to be listed is specified on the FILES> prompt line. The desired listing parameters are specified on the PARAM> prompt line. The parameters which may be specified are as follows: SS:[1] - Starting sample number SL:[1] - Starting line number ES:[NS] - Ending sample number EL:[NL] - Ending line number SINC:[1] - Sample increment LINC:[1] - Line increment PAGE:[] - List total area on one page HEAD:[] - Label to be put on top of printout Note that default values are indicated within the square brackets. As an example, SS=10, LINC=5, HEAD='SAMPLE LISTING' 10 PC-MIPS User Manual would generate a listing of every fifth line and all samples after the first nine. The listing would be titled 'SAMPLE LISTING'. The output file which contains the listed DN values is named LIST.PRT. Sample numbers are given along the top of pages and line numbers are given down the first column of pages. Each page contains 50 lines and 25 samples. Multiple pages are generated until the entire specified area is represented. Example: LIST 9-AUG-84 # of files bit types: 8 1 input max samples: 8000 0 output File name that has the input data to be processed. FILES> sample Valid parameters are: SS:[1] -Starting sample SL:[1] -Starting line ES:[NS] -Ending sample (defaults to the number of samples in the input file) EL:[NL] -Ending line (defaults to the number of lines in the input file) LINC:[1] -Line increment SINC:[1] -Sample increment PAGE:[] -List total area on one page. HEAD:[] -Label to be put on top of listing. (e.g., LINC=1,SINC=1,PAGE, HEAD='LABEL') PARAM> ss=151, es=200, sl=201, el=300, sinc=2 FILES> MAPIMG MAPIMG provides the ability to create an output image which is computed by a user-specified mapping of from 1 to 7 input images. The mapping is done on a pixel-by-pixel basis. The user first specifies the output and input image file names in the normal PC-MIPS format (e.g. out=in1,in2,in3). Next, the user specifies the mapping by entering a mathematical expression which defines the desired mapping. This expression may contain numeric constants, arithmetic operators (+, -, /, *, and **), input image identifiers F1 to F7 which refer to the corresponding input images, and intrinsic functions. Parenthesis may be used to control the order of evaluation. The intrinsic functions supported are as follows: 11 PC-MIPS User Manual Function Returns -------- ------- abs() Absolute value sqrt() Square root log() Base 10 logarithm ten() Power of 10 int() Integer truncation sin() Trigonometric sine cos() '' cosine max() Maximum pixel value min() Minimum pixel value rand() Pseudo-random value in the range 0.0 to 1.0 All functions take a mathematical expression as an argument. This argument is evaluated prior to evaluating the function. For the max(), min(), and mean() functions, the evaluated argument is considered an integer which represents the number of input images to be considered. The sin() and cos() functions assume their argument to represent degrees. The rand() function assumes its argument to be a non-zero seed (or starting value) to be used in generating the random numbers. Example: MAPIMG 10-MAY-88 # of files bit types: 8 1 - 7 input max samples: 3000 1 output File name where the processed results will be stored. (e.g., OUTPUT File name(s) that has (have) the input data to be processed. =INPUT1,INPUT2,INPUT3) FILES> map1 = sample, m1, m2 Enter mapping as a function of the input files F1 - F7 (i.e. 0.33*(F1+F2+F3) or (F1+F2)/F3+127 ) MAPPING> (f1 + 0.5*(max(3) - min(3)) + f2 + f3)/3 FILES> Do you want to display map1 (y/n)? n MAPLN1 MAPLN1 enhances linear features using a semi-linear line detecting/mapping technique. The semi-linear technique is one proposed by G. J. Vanderbrug (1975) and is a compromise between linear and nonlinear detectors. It makes separate comparisons to the left and to the right so it does not respond to edges but will respond to isolated noise. For more information see Bureau of Mines report on automatic linear feature mapping by Chavez (1983). The only user input is the input/output file names at the FILES> prompt. 12 PC-MIPS User Manual Example: MAPLN1 24-JAN-85 # of files bit types: 8 1 input max samples: 3000 1 output File name where the processed results will be stored. (e.g., OUTPUT File name that has the input data to be processed. =INPUT) FILES> linout = sample FILES> Do you want to display linout (y/n)? n MDGRAD This program uses a modified gradient method for linear feature enhancement/mapping, but instead of multiplying the constants A, B, C, and D together they are added. The reason for this is that if only one of them is zero, no information is seen and the computation is faster. The formulas used are: A = |X(i+2,j+1)-X(i+3,j+2)| + |X(i+2,j+2)-X(i+3,j+1)| B = |X(i+1, j )-X(i+4,j+3)| + |X(i+1,j+3)-X(i+4, j )| C = |X(i+2, j )-X(i+3,j+3)| + |X(i+1,j+2)-X(i+4,j+1)| D = |X(i+3, j )-X(i+2,j+3)| + |X(i+1,j+1)-X(i+4,j+2)| Y(i,j) = A + B + C + D where X is the input image and Y is the output image. The only user input is the input/output file names at the FILES> prompt. Example: MDGRAD 21-AUG-84 # of files bit types: 8 1 input max samples: 3000 1 output File name where the processed results will be stored. (e.g., OUTPUT File name that has the input data to be processed. =INPUT) FILES> mdout = sample FILES> Do you want to display mdout (y/n)? n MEDIAN This program applies a 3 x 3 median filter to an image. It will smooth the data in a slightly different manner than that done by a simple average. The median is equal to the middle item 13 PC-MIPS User Manual in a distribution which has been arranged by increasing value. Therefore the median location is equal to the pixel in the exact center and its DN value is assigned to the output pixel. Output and input file names are specified on the FILES> prompt line. There are no other program parameters to be entered. Example: MEDIAN 31-MAY-85 # of files bit types: 8 1 input max samples: 3000 1 output File name where the results are to be stored (e.g., OUTPUT File name where the input data is located = INPUT) FILES> m1=sample FILES> Do you want to display m1 (y/n)? n OVRLAY OVRLAY takes from 2 to 6 input images and generates one output image that indicates which of the input images have data in user specified ranges for the input images being used. This program can be used as part of an IMAGE or GIS processing system. In the GIS mode it can be used to overlay and merge various data sets together and in the IMAGE processing mode as a quick parallelepiped classifier. Currently there is a limit of six input files, but more can be processed by executing the program twice on two different groups of input files and then executing it a third time using the results of the first two groups as input. But this will be good ONLY to map where the data is in given ranges for ALL the input files used. The range of the third image results will be 0 to 3, where 3 indicates where the data is in the given ranges for all the files used. The program uses the MIN and MAX values for each input file supplied by the user to create a look-up table (ITBLA) indexed by the DN values of the input data. The values assigned to ITBLA are either 1 or 2 (1 implies DN value outside range and 2 implies DN value inside MIN-MAX range). This is like stretching all data outside the given range to 1 and inside the range to 2. The values (1 or 2) in this table (ITBLA) are then used to index into a second table (ITBLB) which has the correct sum or DN value for the given input file ranges. The values in the second table (ITBLB) are computed as if the input files were summed after they were given the values of 0 and 1,2,4,8,16, or 32 according to the data being outside (0) or inside (FILE1=1, FILE2=2, FILE3=4, FILE4=8, FILE5=16, and FILE6=32) the specified 14 PC-MIPS User Manual ranges. The data range of the output image will be from zero to (2**NFILES-1) where NFILES = number of input files used (2 to 6). In the case of four input files, the following table (where YES indicates value of given input file is in specified range) gives the value of the output file for possible combinations of input values: INPUT INPUT INPUT INPUT OUTPUT FILE1 FILE2 FILE3 FILE4 DN ----- ----- ----- ----- ------ NO NO NO NO 0 YES NO NO NO 1 NO YES NO NO 2 YES YES NO NO 3 NO NO YES NO 4 YES NO YES NO 5 NO YES YES NO 6 YES YES YES NO 7 NO NO NO YES 8 YES NO NO YES 9 NO YES NO YES 10 YES YES NO YES 11 NO NO YES YES 12 YES NO YES YES 13 NO YES YES YES 14 YES YES YES YES 15 These values may be summarized as follows: 0 = All 4 files NO 1,2,4,8 = Only 1 file YES 3,5,6,9,10,12 = 2 files YES 7,11,13,14 = 3 files YES 15 = All 4 files YES Example: OVRLAY 3-MAY-84 # of files bit types: 8 2 - 6 input max samples: 3000 1 output File name where the processed results will be stored. (e.g., OUTPUT File names that have the input data to be processed. =INPUT1,INPUT2,INPUT3) FILES> overlay = sample, bathy, dpth Valid parameters are: MIN* = Minimum value for FILE(1-6) MAX* = Maximum value for FILE(1-6) * .. Replace with the file number. There are no default values. Values for MIN* and MAX* must be between 0 and 255. (e.g., MIN1=5,MAX1=70,MIN2=22,MAX2=33,MIN3=0,MAX3=96) 15 PC-MIPS User Manual PARAM> min1=3,max1=240, min2=5,max2=170, min3=10,max3=210 FILES> Do you want to display overlay (y/n)? n RATIO This program is used to generate output images which are ratios of input images. Up to four input images and three output images may be specified. The names of the output and input images are specified on the FILES> prompt line. The program then requests ratio parameters on the PARAMS> prompt line. The parameters which may be specified are as follows: R1 - A two digit number specifying how the first output image is to be calculated. The first digit gives the number of the input file which is to be used as the numerator. The second digit gives the number of the input file which is to be used as the denominator. For example R1=13, specifies that the first output image is the ratio of the first and third input images. R2 - Same as R1, but for the second output image. R3 - Same as R1, but for the third output image. FACT - A value which is to be used as a multiplication factor. The default value is 1. Example: RATIO 28-SEP-83 # of files bit types: 8 2 - 4 input max samples: 3000 1 - 3 output File names where the processed results will be stored (1-3 images). (e.g., OUTPUT File names that have the input data to be processed (2-4 images). = INPUT1,INPUT2) FILES> band17,band24=band1,band2,band4,band7 Valid parameters are: R1:[] -First ratio R2:[] -Second ratio R3:[] -Third ratio FACT:[1.0]-Ratio multiplying factor (e.g.,R1=12,R2=13,FACT=1.2) R1=12 means that the 1'st ratio will be input file 1 divided by file 2 R2=13 means that the 2'nd ratio will be input file 1 divided by file 3 R3=23 means that the 3'rd ratio will be input file 2 divided by file 3 16 PC-MIPS User Manual PARAM> r1=14, r2=23 FILES> Do you want to display band17 (y/n)? n ROI This program requires VGA graphics capability and a serial MS-compatible mouse. It allows the user to delineate a sub-image which is extracted from up to 7 input images. The sub-image is determined by entering polygon vertices using a mouse. At the FILES> prompt the user enters output and input file names. The first input file is displayed up to a maximum of 480 lines and 640 samples. Next, the user positions the graphics cursor (an arrow) to an initial vertex and "clicks" on that point by pressing the leftmost mouse button. That point is illuminated and the user moves to the second vertex and "clicks" on it using the same button. A line is drawn connecting the two points. The remaining vertices, up to the last, are specified in the same manner. When the cursor has been positioned on the last vertex, the polygon is closed by "clicking" with the second mouse button. After the polygon has been entered, the program creates the specified sub-image from each of the input files and creates an ascii file named POLYGON.DAT which contains the polygon vertices in a format acceptable to EXT_ROI which can be executed if sub- images from additional files are desired. ROTATE This program rotates an image 90 degrees (i.e., interchanges lines and samples) in the clockwise or counter-clockwise direction. The user specifies output and input file names on the FILES> prompt line. The program, then, requests the direction of rotation. Entering 0 (zero) indicates clockwise and 1 (one) indicates counter-clockwise. If the input image is small enough to reside in the available computer memory, the execution is very fast. For larger images, each pixel must be accessed individually and the execution time is quite slow. If this situation exists, a warning is given. Example: ROTATE 18-OCT-88 # of files bit types: 8 1 input max samples: 3000 1 output File name where the results are to be stored (e.g., OUTPUT File name where the input data is located 17 PC-MIPS User Manual = INPUT) FILES> r1=m1 Enter 0 for clockwise or 1 for counter-clockwise: 0 FILES> Do you want to display r1 (y/n)? n SALTP This program is used to remove bit errors from images. The removal is based on a DN tolerance value which may be specified by the user or determined automatically based on the image histogram. A 3-line by 5-sample box is moved from left to right and from top to bottom through the input image to remove bad pixels from the image. This box is illustrated below: ---- ---- ---- ---- ---- | U1 | U2 | U3 | U4 | U5 | ---- ---- ---- ---- ---- | M1 | M2 | M3 | M4 | M5 | ---- ---- ---- ---- ---- | L1 | L2 | L3 | L4 | L5 | ---- ---- ---- ---- ---- Each time the box is moved, the pixel element in the middle (M3) of the box is compared with averages of other pixels inside the box to determine if the pixel needs to be replaced. This is done in two steps. Step 1 finds the average of the two points at the upper left corner of the box (U1 and U2) and the average of the two points at the lower right hand corner of the box (L4 and L5). If the difference of the two averages is greater than four times the tolerance limit then the box is currently over an edge in the image. If the edge test succeeds (box is over an edge) then the pixel in the middle cannot be corrected. If the edge test fails then step 2 is performed. Step 2 finds the average of the line above (U1-U5) and the line below (L1-L5) the line containing the pixel to be corrected. If the difference of the point in the middle and the average is greater than the acceptable tolerance then the point is considered bad data and is replaced with the average of the box. Example: SALTP 21-DEC-84 # of files bit types: 8 1 input max samples: 3000 1 output File name where the processed results will be stored. (e.g., OUTPUT File name that has the input data to be processed. =INPUT) FILES> cleaned = sample Valid parameters are: 18 PC-MIPS User Manual TOL:[] - Integer DN tolerance value to be used to identify pixels that have a bit error. If not specified, TOL is calculated based on the histogram. PARAM> tol = 11 TOLERANCE= 11 Percentage of pixels changed = 25.8 FILES> Do you want to display cleaned (y/n)? n SHAD2 SHAD2 computes the average of each sample position using every other line. The computed averages are then normalized by either the average value of all the averages or by a user supplied DN value (e.g., NORM). It can be used to correct for shading problems that exist in the horizontal direction (e.g., on sonar or radar data). SHAD2 uses the normalized average values as shading coefficients for image shading corrections. The user specifies values of TOL1 and TOL2 which are used to compute the average values using only data greater than or equal to TOL1 and less than or equal to TOL2. It passes through the input data twice, once to compute the correction coefficients and the second time to apply them. The user may specify a beginning sample number (e.g., ISS) for the computation. The user-specified keywords are as follows: TOL1:[0] - Lower DN limit tolerance to be used to compute averages. TOL2:[maxdn] - High DN limit tolerance to be to compute averages. ISS:[1] - Starting sample user wants to start processing the input data. Output lines will still contain all samples. NORM:[ave] - DN values to be used to normalize all the averages. Default is the average of all the averages. Example: SHAD2 27-APR-84 # of files bit types: 8 1 input max samples: 3000 1 output File name where the processed results will be stored. (e.g., OUTPUT File name that has input data to be processed. =INPUT) FILES> shaded = sample Valid parameters are: TOL1:[0] - Lower DN tolerance to be used to compute averages. TOL2:[maxdn] - High DN tolerance to use to compute averages. 19 PC-MIPS User Manual ISS:[1] - Starting sample to start processing. Output lines will still contain all samples. NORM:[ave] - DN value to be used to normalize all the averages. Default is average of averages. (e.g., TOL1=1,TOL2=250,ISS=17,NORM=100) PARAM> tol1=5, tol2=250, norm=100 FILES> Do you want to display shaded (y/n)? n STRTCH The STRTCH program performs contrast stretching on images. The user can specify up to ten stretch intervals. A stretch interval is specified by four values. These four values are given in terms of DN values. They are IN_LOW, IN_HIGH, OUT_1, and OUT_2. IN_LOW and IN_HIGH represent an input interval. OUT_1 and OUT_2 represent the corresponding output interval. Input DN values are assigned to values in the output interval based on their position within the input interval. If IN_LOW is equal to IN_HIGH, input DN values of IN_LOW are assigned the value of OUT_1. IN_HIGH cannot be less than IN_LOW. Input DN values not within any of the specified input intervals are unchanged. If OUT_1 is equal to OUT_2, all DN values in the input interval are reassigned a value of OUT_1. If OUT_2 is less than OUT_1, an 'inverse' stretch is performed (i.e., increasing input DN values are reassigned to decreasing output DN values). Note: STRTCH differs from the MIPS STRETC program which uses stretch 'pairs' instead of stretch 'intervals' to specify the stretch operation. Example: STRETC 09-NOV-88 # of files bit types: 8 1 input max samples: 3000 1 output File name where the processed results will be stored. (e.g., OUTPUT.EXT File name that has the input data to be processed. =INPUT.EXT) FILES> result = image Enter number of stretch intervals: 4 Enter IN_LOW, IN_HIGH, OUT_1, OUT_2 for interval 1: 0, 15, 0, 0 Enter IN_LOW, IN_HIGH, OUT_1, OUT_2 for interval 2: 16, 71, 0, 127 Enter IN_LOW, IN_HIGH, OUT_1, OUT_2 for interval 3: 72, 127, 128, 255 20 PC-MIPS User Manual Enter IN_LOW, IN_HIGH, OUT_1, OUT_2 for interval 4: 128, 255, 255, 255 FILES> Do you want to display result (y/n)? n SUBIMG The sub-image program is used to extract a user-specified sub-image. This program is similar to the MIPS DK2DK program. Additionally, the output image may be 'flipped' (or 'reversed') in the line and/or sample directions. The user is prompted for an input file name. After the input file name is specified, the number of lines and samples in the input file are displayed and the user is prompted for starting/ending line and sample numbers. The output file is 'flipped' by specifying an ending value less than the starting value. Finally, the user is prompted for an output file name. Example: SUBIMG (10/88) allows the extraction and/or "flipping" of sub- images. The dimensions of the sub-image are entered by the user. "flipping" is accomplished by indicating starting values greater than ending values. Enter input file name: study study contains 1112 lines and 984 samples Enter starting and ending line numbers: 1, 420 Enter starting and ending sample numbers: 981, 400 Enter output file name: sub_stdy Enter input file name: WTCMBO This program is used to generate an output image which is the weighted sum (i.e., a linear combination) of up to four input images. The output is calculated using the following formula: Y = C1*X1 + C2*X2 + C3*X3 + C4*X4 +C5 where Y is the output image, C1-C5 are user-specified constants, and X1-X4 are user-specified input images. The computation is done using 16-bit arithmetic. Example: WTCMBO 04-JUN-84 # of files bit types: 8 1 - 4 input max samples: 3000 1 output 21 PC-MIPS User Manual File name where the processed results will be stored. (e.g., OUTPUT File name(s) that has (have) the input data to be processed. =INPUT1,INPUT2,INPUT3) FILES> sum = sample, shade Valid parameters are: CON1:[.25] - Weighting coefficient for input file 1. CON2:[.25] - Weighting coefficient for input file 2. CON3:[.25] - Weighting coefficient for input file 3. CON4:[.25] - Weighting coefficient for input file 4. CON5:[0] - Add-back constant for output file. Used to center the results in 8-bit positive range. (e.g., CON1=1.0,CON2=-1.0,CON3=.5,CON5=100) PARAM> con1=0.75, con2=1.25, con5 = -127 FILES> Do you want to display sum (y/n)? n 22