文件列表:

Makefile.in (2033, 2002-07-27)
configure (70085, 2002-06-22)
configure.in (448, 2002-06-18)
correct.c (21710, 2004-08-13)
correct.h (792, 2003-08-09)
eorc.h (1847, 2002-07-23)
getecamimg.cpp (29645, 2004-08-19)
install-sh (5598, 2001-06-21)
iqeye.cpp (34151, 2004-08-19)
iqeye.h (7726, 2003-08-09)
iqshow.cpp (13604, 2003-08-09)
iqshow.mak (1737, 2002-08-08)
makebor.bat (219, 2003-08-09)
makefile.msvc (710, 2004-07-13)
makemsvc.bat (117, 2003-08-09)
makeoid.bat (161, 2003-08-09)
md5c.c (9920, 2001-03-07)
mymd5.h (1454, 2001-03-07)
oid.cpp (1685, 2002-07-27)
sex.h (454, 1999-05-17)
streambf.cpp (249, 2001-08-15)
streambf.h (791, 2001-08-15)
udpcmn.h (11497, 2002-06-22)
x.cpp (3857, 2003-08-09)
x.h (445, 2003-08-09)
msvc (0, 2004-08-19)
msvc\iqshow.dsp (3754, 2003-08-09)
msvc\iqshow.dsw (549, 2003-08-09)

Readme file for the getecamimg sample application for the IQeye. What is getecamimg? ------------------- getecamimg is a sample command line application, compileable under *nix or Windows, which demonstrates the use of the CIQeye class to request raw images from the IQeye camera. The image data is transferred in raw form, as opposed to getting a JPEG image with a browser. what is iqshow? --------------- iqshow also demonstrates the CIQeye class, but displays the resulting images. It also compiles under *nix and Windows. Building under *nix ------------------- ./configure make all will build getecamimg under *nix systems make iqshow will build iqshow under *nix systems Building under Windows ---------------------- We have sample .bat files which show how to compile under borland and msvc compilers. They're simple one-liners. The important things to note are that you have to compile with -DWINDOWS and include the winsock2 library. iqshow.mak is a Borland makefile for iqshow. The CIQeye Class ---------------- CIQeye is a C++ class which wraps the nview protocol for speaking to an IQeye camera. It was written to provide an easy interface for grabbing raw images from the camera, so folks don't have to plow through the nview protocol specification. getecamimg.cpp demonstrates the use of this class. It also demonstrates how to convert from the raw image to RGB data (the PPM format). Why use this class? Because this is the only way to get raw image data from the camera. If you go over HTTP, you can only get JPEGs. Quick Start Usage ----------------- getecamimg -bin > img.raw Will grab a raw image from and write it to file 'img.raw'. getecamimg -ppm > img.ppm Will grab a raw image from and write it out in PPM (RGB) format to file 'img.ppm'. Image Format ------------ The nview protocol grabs raw image data from the camera. Each byte represents either a R, G, or B value, straight from the imager. If you want to create an RGB image from this (PPM for instance), then you'll need to convert it using the color conversion matrix which gets sent with each frame. This matrix is the same one that the camera uses when it generates JPEG images, so JPEGs from the camera and PPMs from getecamimg should look about the same. See DoConvert() in correct.c for how to do the conversion. Note that the code makes no assumptions about how the Bayer pattern comes out of the camera - all the information about which values are actually R, G, or B is encoded in the conversion matrix. Using the API ------------- Basic usage: int rv; int imglen; unsigned char *bf = (unsigned char *)malloc(MAX_COLS * MAX_ROWS); CIQeye *iq = new CIQeye (host, "root", "system"); rv = iq->readImage (iq->SubXorg, iq->SubYorg, iq->SubWidth, iq->SubHeight, bf, &imglen); delete iq; free (bf); The constructor opens up a socket. The default priviledged username and password is "root" and "system." readImage() grabs a raw frame with the given window coordinates into the user-supplied buffer. The Subxxx variables are correspond to the default crop window on the camera. After you grab the image, you can do anything you want with it, then call the destructor to clean up. CIQeye member functions ----------------------- o CIQeye (char *ip_name, char *user = NULL, char *passwd = NULL) Constructor for the CIQeye class. ip_name: ip address or hostname of camera user: username to authenticate with ("root" to authenticate with system privileges) password: password to authenticate with ("system" is the default system password) This function initializes the networking, gets authentication credentials, image options, and image configuration from the camera. o ~CIQeye () Class destructor. Cleans everything up. o bool readImage (int clipx, int clipy, int clipw, int cliph, StreamBf *imgbf, int *imglen, bool domask = false, bool compressed = false, bool correct = false); Reads one image from the camera in raw bayer pattern form. clipx: starting column clipy: starting row clipw: width clipy: height imgbf: buffer for image - needs to be allocated to max possible image size by caller imglen: receives total bytes in image domask: true if we want a masked image compressed: true if we want a compressed image correct: true to enable hotpixel correction Returns true on success, false on failure o bool getMatrix (long *matrix) Gets the EORC matrix from the camera. The EORC matrix is what allows you to convert from bayer pattern to RGB. matrix: array of longs to receive the matrix (must be EORCMTXCNT long) Returns true on success, false on failure o bool setImageMask (int start_row, int num_rows, int *starts, int *lengths) Send an image mask to the camera. start_row: starting row of this mask (base 0) num_rows: number of rows in this mask starts: array of ints which specify starting column (base 0) lengths: array of ints which specify row length for each start Returns true on success, false on failure Setting an image mask is a way to get an arbitrary blob from the camera instead of a rectangle. The protocol supports blobs with holes in it, but this function does not. So you can specify a disk but not a doughnut. If you set domask to true when you readImage(), the resulting image buffer will consist of only the pixels which match the masked area o bool clearImageMask () Tells the camera to clear the image mask. You always want to do this as part of your cleanup procedure since the image mask takes up memory on the camera. And if you don't clear it, it will stick around for the next guy. Returns true on success, false on failure o void setFlip (unsigned char flip) Tells the camera to flip/rotate the image as per the flags in udpcmn.h o void setDS (unsigned char ds) Sets the downsample factor. The image gets downsample in each dimension by ds. o void setSharpen (unsigned short sharpen) Sets the sharpening factor on the camera. The camera does not sharpen for you - this just affects the EORC sharpening values for when you sharpen locally. o void setSharpNoI (bool val) Tells the camera to send sharpen values for a non-interpolated image (if val == true) or an interpolated image (val == false). By default, the camera sends sharpening values assuming that you will interpolate. If you are going to us a non-interpolate color correction, you should ask for the non-interpolate sharpening scalars. o void setExtraProc (unsigned char extra) This sets the value that is put in the extra processing field of the image request. Use this if you have custom firmware which registers an extra processing hook into the nview protocol (see oem.c in the partner source code). o void setGamma (unsigned char gamma) Sets the gamma value sent with the image request. Currently you have to gamma correct locally, so this doesn't do any good. o void setManualGain (int manual) Determines how setGainPercent() behaves. manual: if 0, turns on auto gain, else turns on manual gain Returns true on success, false on failure o bool setGainPercent (int gain_percent) Sets gain on the camera as a percentage from 0 (lowest) to 100 (highest). If gain is manual, gain in the camera is held to reflect gain_percent. If gain is auto, the target gain value is set to gain_percent gain_percent: percent from 0 to 100 Returns true on success, false on failure o bool getGainPercent (int *gain_percent) Gets the gain from the camera, expressed as a percent from 0 to 100. If gain is manual, this is the hold value. If gain is auto, this is the target value. Returns true on success, false on failure o unsigned long getTimestamp () Returns the timestamp from the last image read. Every tick of this timer is 10ms and is relative, not absolute (wall) time. o unsigned short getMaxWidth () Returns the maximum imager width o unsigned short getMaxHeight () Returns the maximum imager height o unsigned short getLastWidth () Returns the width of the last image read o unsigned short getLastHeight () Return the height of the last image read o bool GetOid (const char *oid, char *val) Get a runtime OID value from the camera. oid: OID to get the value of value: where to put the value OIDs and the OID table are documented elsewhere. Returns true on success, false on failure o bool SetOid (const char *oid, const char *val) Set a runtime OID on the camera oid: OID to set value: new value for the OID Returns true on success, false on failure o const char *strerror (int code) Converts a failure code to a meaningful string. code: FailReason CIQeye member variables ----------------------- o int FailReason Last error code. See iqeye.h for values, or call iq->strerror (iq->FailReason) for a string. o unsigned short SubWidth Width of the camera's crop window o unsigned short SubHeight Height of the camera's crop window o unsigned short SubXorg Left coordinate of the camera's crop window o unsigned short SubYorg Top coordinate of the camera's crop window o unsigned short MaxWidth Maximum window width o unsigned short MaxHeight Maximum window height o unsigned short LastWidth Width of last image o unsigned short LastHeight Height of last image o unsigned char CurGamma Camera's current gamma value o int ExtraLen Length of any meta data delivered with last image o StreamBf *ExtraBf Extra data (if any) delivered with last image

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：