Homepage

Demos

Overview

Downloads

Tutorials

Reference

Credits

RawCameraGenerator.h Go to the documentation of this file. //-*-c++-*- #ifndef INCLUDED_RawCameraGenerator_h_ #define INCLUDED_RawCameraGenerator_h_ #include "Vision/FilterBankGenerator.h" class OFbkImageVectorData; class OFbkImageInfo; //! Generates FilterBankEvents containing raw camera images directly from the system (doesn't make a copy) class RawCameraGenerator : public FilterBankGenerator { public: //! constructor, numRawLayers is the number of real layers passed from the system, numCalcLayers is the total number of layers to make available /*! The extra calculated layers are simply created by giving larger * increments, they reference the same data (no computational cost, * marginal memory costs... it's just nice to have more layers in * the image pyramid. * * However you'll have to write your code to use the getIncrement() * value properly if you want it to work. Otherwise, you could * write a de-interlace generator which will do the resampling. * * The top most layer (largest index) is a double-scale image. For * the Y-channel, it is reconstructed from the 4 Y channels. For * the other channels, it just does a fast scaling (but these * operations do cost, so use the top layer conservatively. * * @see FilterBankGenerator for information on serialization format */ RawCameraGenerator(unsigned int numRawLayers, unsigned int numCalcLayers, EventBase::EventGeneratorID_t gid, unsigned int sid, unsigned int mysid); //! destructor virtual ~RawCameraGenerator(); static std::string getClassDescription() { return "Translates OFbkImageVectorData objects from the system into a slightly more accessible FilterBankEvent for further processing"; } //! holds id values for specifying image channel/bands enum channel_id_t { CHAN_Y, //!< Y (intensity) channel CHAN_U, //!< Cr (approx. pink,red,yellow) channel CHAN_V, //!< Cb (approx. blue,green,yellow) channel CHAN_Y_DY, //!< vertical derivative of Y channel (aka LH) CHAN_Y_DX, //!< horizontal derivative of Y channel (aka HL) CHAN_Y_DXDY, //!< vert. & horiz. derivatives of Y channel (aka HH) NUM_CHANNELS //!< number of channels per resolution layer }; //! called with system's image info virtual void processEvent(const EventBase& event); //! the memory for all layers except the double layer was allocated by system, so we just set them to NULL before calling FilterBankGenerator::freeCaches() so it won't try to delete them virtual void freeCaches(); virtual unsigned int getBinSize() const; virtual unsigned int LoadBuffer(const char buf[], unsigned int len); virtual unsigned int SaveBuffer(char buf[], unsigned int len) const; virtual unsigned int SaveFileStream(FILE* f) const; //!< overrridden to allow saving direct to file without an extra buffer copy protected: //! ohh, this is so yucky. But we need to get around heap allocation bugs with large regions in Aperios :( /*! So we store the double size images statically within the class. * What a waste of memory since we're probably not using these very * much, and they are rather large... :( * * The 2.0 release had the proper dynamic allocation * implementation, but has been replaced in 2.0.1 with this static * allocation so it won't crash. * * Of course, if the memory allocation bug is never fixed, and * camera sizes increase in later models, this will have to be * modified. */ unsigned char dblRes[NUM_CHANNELS][416][320]; virtual void setNumImages(unsigned int nLayers, unsigned int nChannels); virtual void setDimensions(); //!< resets the width, height, skip and stride values virtual unsigned char * createImageCache(unsigned int layer, unsigned int chan) const; virtual void calcImage(unsigned int layer, unsigned int chan) const; virtual void destruct(); //! returns the value to pass to OPENR functions corresponding to the desired channel_id_t. inline static unsigned int mapChannelID(channel_id_t chan); //! Creates a double resolution version of a channel, doesn't do any smoothing void upsampleImage(channel_id_t chan) const; //! Creates a double resolution Y-channel from the Y and Y gradient channels void reconstructImage() const; //! helper functions for reconstructImage() - see documentation for that function static inline unsigned char clipRange(int val) { if (val < 0) { return 0; } else if (val > 255) { return 255; } else { return (unsigned char)val; } } unsigned int numRealLayers; //!< the number of "real" layers being sent from the system unsigned char ** layers; //!< points to the image data supplied by the system, one per layer (all channels are interleaved by row) const OFbkImageInfo ** imageInfos; //!< image info provided by the system private: RawCameraGenerator(const RawCameraGenerator& fbk); //!< don't call const RawCameraGenerator& operator=(const RawCameraGenerator& fbk); //!< don't call }; /*! @file * @brief Describes RawCameraGenerator, which generates FilterBankEvents containing raw camera images * @author ejt (Creator) * * $Author: ejt $ * $Name: tekkotsu-2_0_1+Doc $ * $Revision: 1.13 $ * $State: Exp $ * $Date: 2004/02/05 23:33:48 $ */ #endif

Tekkotsu v2.0.1+Doc

Generated Mon Feb 9 22:16:50 2004 by Doxygen 1.3.5