Files
rtabmap/corelib/include/rtabmap/core/CameraModel.h
T
matlabbe ee49beaf4f Adding doc and tests (#1492)
* added doc and tests for util2d.h

* updated cmake-ros ci

* Added util3d.h doc and tests

* util3d_transforms.h: Added doc and tests

* util3d_filtering.h: started doc and test

* util3d_filtering.h: more tests and doc

* Added more doc/tests

* finished util3d_filtering doc and tests

* added test for util2d::depthBleedingFiltering

* Added util3d_registration tests

* Added util3d_features.h doc/tests

* added doc/tests for util3d_correspondences.h

* added doc/gtest for util3d_mapping.h (missing hpp functions)

* finished testing util3d_mapping.hpp

* Added util3d_motion_estimation.h tests (2D->3D done)

* finished util3d_motion_estimation.h tests

* minimal util3d_surface.h

* Added Transform and VisualWord tests

* Added doc for CameraModel and StereoCameraModel

* Added more logs in ros ci

* Passing tests on fical

* improved all devcontainer

* added devcontainer kilted, fixed source setup.bash, removed ldconfig in ros-cmake workflow

* cleanup

* source ros

* Added utilite tests

* Added testing to appveyor, github actions cancellable on re-commit on same branch

* appveyor testing without all targets

* appveyor: specifying ALL_BUILD target

* Fixed Util2dTest.NMSImageBoundsRespected test

* Fixing PCL Indices error on old pcl

* Added VWDictionary tests and doc. Fixed LSH not working (fix from https://github.com/flann-lib/flann/pull/472

* fixing some appveyor CI errors, added test to check dictionary serialization against all type

* Added StereoDense, StereoBM and StereoSGBM doc and tests

* Added Stereo tests

* Added CameraModel and StereoCameraModel tests

* Added doc and test for Statistics

* Added doc/tests for Signature

* Added doc/test for SensorEvent, added doc for SensorCaptureInfo

* Added doc to SensorData

* Added SensorData tests

* Added SensorCapture and SensorCaptureThread doc and tests

* fixed sensordata test

* updated SSC test and doc

* Added doc and tests for BayesFilter class

* Enabled testing on mac, updated windows testing like on linux

* added test_link

* fixed unresolved on windows

* fixed ThreadHandle error on macos ci

* Added GPS and GeodeticCoords tests

* Added tests for compression

* Added Odometry tests (base class only)

* Added DBDriver tests

* Added coverage report

* uniformized test names

* fixing concurancy and coverage ci

* dont built tools, examples and app for coverage build

* fixed report tool rebuilt without qt compilation error

* updated coverage option

* updated coverage config

* added doc CI job

* fixing windows and mac ci errors

* Added DBDriverSqlite3 tests

* Added IMU tests

* Added Graph tests

* fixing flaky macos test

* Added IMUThread and IMUFilter tests

* Added Landmarks tests

* Added LASWriter tests

* fixing seed flaky test

* fixing flaky macos timing tests

* Added LocalGrid tests

* Added LocalGridMaker tests

* fixing ci errors

* Added GlobalMap tests

* Added doc for EnvSensor

* Added Features2D tests

* Added Registration tests

* Added RegistrationVis tests

* Added doc for Rtabmap and Memory classes

* Added Memory and Rtabmap tests

* making some tests less flaky

* lcov 1.14 support

* updated compatible tool arguments

* Added integration tests (RGB-D, Stereo, Lidar2d, Lidar3d)

* More octomap checks

* Refactored how/when python interpretor is created to simplify library usage

* Added python tests

* fixed some flaky tests

* suppressed some third party related warnings

* fixed ceres tests

* more flaky fixes

* Fixing tests without libpointmatcher

* Added RANSAC rejection filter to PCL ICP

* fixing multi platform flakiness

* Added test to detect regression

* Fixing windows pcl link error

* fixed some macos flakiness

* bigger 2D2D registration error on opencv 4.6.0

* flakiness

* fixing flaky tests on windows and mac

* flaky thread test on slow mac VM

* windows slow test

* fixing more ci erros

* fxing temp dir on windows

* Added Optimizer tests and discovered some bugs (fixed)

* fixing flaky tests in mac and windows

* Added Optimizer doc

* Added GTSAM BA, updated Ceres to use g2o ba parameters. Renamed g2o's ba related parameters to Optimizer group and used by both gtsam and ceres.

* fixing build without gtsam

* fixing home dir

* fixing python ci isssues

* Added multicam ba tests

* Added Ceres multicam BA support

* Aligned BundleAdjustment parameters with Optimizer/Strategy to avoid confusion in the code

* Added BA integration test

* Added robust graph optimization integration test

* Added loop3it test

* Added stereo20Hz test

* Added smartfactor gtsam

* Fixed bugged check and warn if python didn't return any descriptors

* Fixing gtsam version build issues

* fixing tilt on windows ci

* loosing ceres integration test for ci

* mac ci flakiness

* updating missing param in gui

* updating test bound for mac

* added appearance-based tests, set min gftt quality to quality level

* testing more stuff

* improving features2d tests

* ci flakiness

* fixing flaky ci

* ci fixes

* flaky fixes

* Added RegistrationIcp tests

* Added icp integration test with real-worl corridor like env

* intermediate nodes

* fixing enum

* Updated test to catch #1714

* Fixed 2d corridor failing on pcl

* flaky pnp test

* flaky brisk test

* Set rtabmap_integration test as long

* updating loop closure test

* flaky ci tests

* TEsting roundtrip g2o/toro save/load

* loosing test bound

* fixed cuda capable checks

* flaky tests

* Debugging test hanging

* more debugging stuff

* updating limit

* windows: disabled cuda on ci to avoid incompatible driver issue. Fixing a bad test mem allocation

* trying fixing cuda hanging issue

* fixing ci flakyness

* flaky tests

* Updated BOW flaky tests by checking min precision/recall instead of recall@100precision. Fixed signature test

* CameraModel::load() test initRectificationMap param

* test dbdriver load dictionary idsOnly

* Memory: test keepLinkedInDb param

* added dummyDictionary tests

* test intermediate nodes count

* Added MarkerDetector tests

* reverted breaking change of UMutex and USemaphore

* Features2d: fixed compiltion warnings with clang about override

* clang warnings

* fixing test build with pcl 1.8

* g2o and gtsam build errors on android

* opencv5 test fixes

* disabled testing for ios and android builds

* normalized endline characters for easier diff

* added LF CRLF rule

* bump 0.23.10. fixing doc version

* Publish rtabmap website doc from ci

* fixing MSCVC build error

* macos icp flaky test

* fixing ceres macos test bound

* ficing more flaky tests

* fixing opencv5 related test errors. Also fixed an actual bug in ENU_WGS84ToGeocentric_WGS84()

* added comment about mrpt change

* removed rosdoc2 (will add it for rtabmap_ros later)

* fixing website style

* updated download links

* locally deployable website with api

* sweep doxygen issues

* improved/revised doxygen main pages

* removed examples empty page

* Updated doxygen style

* more concise doxygen groups

* added api link on main readme

* fixing utilite test error

* fixing CommonFilteringGroundNormalsUp test

* updated precisionRecall test bounds for Freak and brief descriptors

* fixing scale check in ba tests

* disabled tests on windows cuda build (missing dlls amd runner cannot test cuda anyway)

* ceres: missing suitesparse dep in windows ci

* adjusting recall thr for fast/freak

* ficing more flaky tests

* fixing flaky tests

* disabled coverage in ros ci

* Enable integration tests for ros ci jobs

* loosing up some threshold for failing tests

* trigger cache

* fixing test data in ros ci. Updated flaky test for mac

* slaking some test limit

* Fixed rtabmap-detectMoreLoopClosures inverted output value

* loosing up sift recall on mac

* optimizer re-ordered distribution for reproducible results (mac g2o)

* macos dump test crash log

* combining all tests to save time on shared library reload. Also fixed Logs with missing arguments.

* Added ENABLE_FORMAT_ERRORS cmake option

* do test only one time

* fixed all format warnings

* format security android build errors

* less verbose tests

* updated ImuUThread test

* fixed a log

* Fixed libpointmatcher 2d normals eigen issue

* Fixing libpointmatcher conversion issues

* fixing libpointmatcher test on windows ci

* cleanup comments, relax some test thr

* disabled sequoia-intel ci build (too flaky, would need extensive testing directly on that machine)
2026-08-06 13:32:20 -07:00

594 lines
25 KiB
C++
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
/*
Copyright (c) 2010-2016, Mathieu Labbe - IntRoLab - Universite de Sherbrooke
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of the Universite de Sherbrooke nor the
names of its contributors may be used to endorse or promote products
derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef CAMERAMODEL_H_
#define CAMERAMODEL_H_
#include <opencv2/opencv.hpp>
#include "rtabmap/core/rtabmap_core_export.h" // DLL export/import defines
#include "rtabmap/core/Transform.h"
namespace rtabmap {
/**
* @class CameraModel
* @brief Represents a pinhole camera model containing intrinsic and extrinsic
* parameters, used for projection, rectification, and transformation.
*
* This class encapsulates camera calibration data, including intrinsic parameters (fx, fy, cx, cy),
* distortion coefficients, rectification and projection matrices. It provides utility functions
* for image rectification, projection from 2D to 3D, and vice versa.
*
* This class supports the 4 to 14 parameters Radial Tangential distortion model (also called Plumb Bob or
* Brown-Conrady model) and 4 parameters Fish Eye model (also known as Equidistant model).
*
* @see OpenCV's calib3d module for all supported camera models.
*/
class RTABMAP_CORE_EXPORT CameraModel
{
public:
/**
* @brief Returns the default optical rotation to convert image coordinates to robot coordinates.
*
* Image frame: x -> right, y -> down, z -> forward
* Robot frame: x -> forward, y -> left, z -> up
*
* @return Transform rotation matrix.
*/
static Transform opticalRotation() {return Transform(0,0,1,0, -1,0,0,0, 0,-1,0,0);}
public:
/// Default constructor.
CameraModel();
/**
* @brief Constructor using full camera parameters.
*
* @param name Camera name or ID.
* @param imageSize Size of the image (width x height).
* @param K Intrinsic matrix (3x3).
* @param D Distortion coefficients, 1xN matrix where N is between 4 and 14 parameters:
* k1,k2,p1,p2[,k3[,k4,k5,k6[,s1,s2,s3,s4[,tx,ty]]]]).
* To set Fish Eye / Equidistant model, it is implicitly used if we
* provide 6 values like this: [k1,k2,0,0,k3,k4], where you only need to
* fill "k" parameters.
* @param R Rectification matrix (3x3).
* @param P Projection matrix (3x4).
* @param localTransform Local transform to apply to the camera frame.
*/
CameraModel(
const std::string & name,
const cv::Size & imageSize,
const cv::Mat & K,
const cv::Mat & D,
const cv::Mat & R,
const cv::Mat & P,
const Transform & localTransform = opticalRotation());
/**
* @brief Minimal constructor using intrinsic parameters. This assumes the images are already rectified.
*
* @param fx Focal length x.
* @param fy Focal length y.
* @param cx Principal point x.
* @param cy Principal point y.
* @param localTransform Local transform to apply to the camera frame.
* @param Tx Baseline * fx (optional). Mainly used in case of stereo pair.
* @param imageSize Image size (optional).
*/
CameraModel(
double fx,
double fy,
double cx,
double cy,
const Transform & localTransform = opticalRotation(),
double Tx = 0.0f,
const cv::Size & imageSize = cv::Size(0,0));
/**
* @brief Minimal constructor with name for saving.
*/
CameraModel(
const std::string & name,
double fx,
double fy,
double cx,
double cy,
const Transform & localTransform = opticalRotation(),
double Tx = 0.0f,
const cv::Size & imageSize = cv::Size(0,0));
/// Destructor.
virtual ~CameraModel() {}
/**
* @brief Initializes the rectification maps used to undistort and rectify images.
*
* This function prepares the `mapX_` and `mapY_` lookup tables used for image rectification.
* It supports both standard radial-tangential distortion and fisheye/equidistant distortion models.
*
* - If the distortion model is **fisheye** (indicated by `D_.cols == 6`), it uses
* `cv::fisheye::initUndistortRectifyMap()` to create the rectification maps. This requires OpenCV ≥ 2.4.10.
* - Otherwise, it uses the standard `cv::initUndistortRectifyMap()` for plumb bob or rational polynomial models.
*
* @pre The camera model must be valid for rectification:
* - `imageSize_` must be non-zero.
* - `D_` must be a 1-row matrix with an accepted number of columns (4, 5, 6, 8, 12, or 14).
* - `R_` must be a 3x3 rectification matrix.
* - `P_` must be a 3x4 projection matrix.
*
* @return `true` if the rectification maps were successfully initialized (`mapX_` and `mapY_` are not empty),
* `false` otherwise.
*
* @see isRectificationMapInitialized(), rectifyImage(), rectifyDepth()
*
* @warning Requires OpenCV 2.4.10 or newer for fisheye support. If the version is older, fisheye rectification will not work.
*/
bool initRectificationMap();
/**
* @brief Checks if the rectification map has been initialized.
* @return True if both mapX_ and mapY_ are initialized.
*/
bool isRectificationMapInitialized() const {return !mapX_.empty() && !mapY_.empty();}
/**
* @brief Checks if the model is valid for 2D->3D projection.
*/
bool isValidForProjection() const {return fx()>0.0 && fy()>0.0 && cx()>0.0 && cy()>0.0;}
/**
* @brief Checks if the model is valid for 3D->2D reprojection.
*/
bool isValidForReprojection() const {return fx()>0.0 && fy()>0.0 && cx()>0.0 && cy()>0.0 && imageWidth()>0 && imageHeight()>0;}
/**
* @brief Checks if the model has sufficient data for image rectification.
*/
bool isValidForRectification() const
{
return imageSize_.width>0 &&
imageSize_.height>0 &&
!K_.empty() &&
!D_.empty() &&
!R_.empty() &&
!P_.empty();
}
/// Sets the camera name, used to set a camera name when saving to a file.
void setName(const std::string & name) {name_=name;}
/// Returns the camera name.
const std::string & name() const {return name_;}
/// Returns focal length in x.
double fx() const {return P_.empty()?K_.empty()?0.0:K_.at<double>(0,0):P_.at<double>(0,0);}
/// Returns focal length in y.
double fy() const {return P_.empty()?K_.empty()?0.0:K_.at<double>(1,1):P_.at<double>(1,1);}
/// Returns principal point x.
double cx() const {return P_.empty()?K_.empty()?0.0:K_.at<double>(0,2):P_.at<double>(0,2);}
/// Returns principal point y.
double cy() const {return P_.empty()?K_.empty()?0.0:K_.at<double>(1,2):P_.at<double>(1,2);}
/// Returns the x translation (usually fx * baseline in case of stereo, otherwise would be 0).
double Tx() const {return P_.empty()?0.0:P_.at<double>(0,3);}
/// Returns the raw intrinsic matrix (before rectification).
cv::Mat K_raw() const {return K_;}
/// Returns the raw distortion coefficients (before rectification).
cv::Mat D_raw() const {return D_;}
/// Returns the rectified camera intrinsic matrix if the projection matrix P exists, otherwise returns the raw intrinsic matrix.
cv::Mat K() const {return !P_.empty()?P_.colRange(0,3):K_;}
/// Returns the rectified distortion coefficients (1x5 filled with zeros) if the projection matrix P exists, otherwise returns the raw distortion coefficients.
cv::Mat D() const {return P_.empty()&&!D_.empty()?D_:cv::Mat::zeros(1,5,CV_64FC1);} // if P exists, return rectified version
/// Returns the rectification matrix.
cv::Mat R() const {return R_;}
/// Returns the projection matrix.
cv::Mat P() const {return P_;}
/// Sets the local transform of the camera (base frame to optical frame).
void setLocalTransform(const Transform & transform) {localTransform_ = transform;}
/// Returns the local transform (base frame to optical frame).
const Transform & localTransform() const {return localTransform_;}
/**
* @brief Sets the image size of the camera model and updates the principal point if undefined.
*
* This function updates the internal image size (`imageSize_`) with the provided size.
* If the intrinsic matrices (`K_` or `P_`) are present and the principal point coordinates
* (`cx`, `cy`) are zero, they are set to the image center (`width/2 - 0.5`, `height/2 - 0.5`).
*
* This ensures the camera model remains valid and useful even when the calibration file
* has no principal point set or the image size is updated manually.
*
* @param size The new image size. It must be either both dimensions zero (clearing) or both positive.
*
* @pre `size.width > 0 && size.height > 0` or `size.width == 0 && size.height == 0`
* @post Updates the `imageSize_`, and adjusts `cx` and `cy` in `K_` and `P_` if they were initially zero.
*
* @warning If `K_` or `P_` are not initialized (`empty()`), no updates will be applied to them.
*
* @see imageSize(), imageWidth(), imageHeight()
*/
void setImageSize(const cv::Size & size);
const cv::Size & imageSize() const {return imageSize_;}
int imageWidth() const {return imageSize_.width;}
int imageHeight() const {return imageSize_.height;}
/**
* @brief Returns the horizontal field of view (FoV) in radians.
*
* The FoV is computed using the pinhole camera model as:
* \f[
* \text{FoV}_x = 2 \cdot \tan^{-1}\left(\frac{\text{image width}}{2 \cdot f_x}\right)
* \f]
*
* @return Horizontal field of view in radians. Returns 0.0 if image width or focal length is invalid.
*/
double fovX() const;
/**
* @brief Returns the vertical field of view (FoV) in radians.
*
* The FoV is computed using the pinhole camera model as:
* \f[
* \text{FoV}_y = 2 \cdot \tan^{-1}\left(\frac{\text{image height}}{2 \cdot f_y}\right)
* \f]
*
* @return Vertical field of view in radians. Returns 0.0 if image height or focal length is invalid.
*/
double fovY() const;
/**
* @brief Returns the horizontal field of view in degrees.
*
* Converts the result of `fovX()` from radians to degrees.
*
* @return Horizontal field of view in degrees. Returns 0.0 if the result is invalid.
*/
double horizontalFOV() const;
/**
* @brief Returns the vertical field of view in degrees.
*
* Converts the result of `fovY()` from radians to degrees.
*
* @return Vertical field of view in degrees. Returns 0.0 if the result is invalid.
*/
double verticalFOV() const;
/// Checks if the distortion model is fisheye (6 coefficients: k1,k2,0,0,k3,k4).
bool isFisheye() const {return D_.cols == 6;}
/**
* @brief Loads the camera model parameters from a YAML calibration file.
*
* This method attempts to read camera intrinsic/extrinsic parameters and image size from a YAML file,
* typically in the ROS calibration format. If the distortion model is "fisheye" or "equidistant", we expect
* 4 coefficients, which are converted to a 6-coefficient format for internal representation.
*
* Fields loaded (if present):
* - `camera_name`
* - `image_width` (pixels)
* - `image_height` (pixels)
* - `camera_matrix` (K, 3x3 double matrix)
* - `distortion_coefficients` (D, 1xN double matrix)
* - `distortion_model` (string: name of the model)
* - `rectification_matrix` (R, 3x3 double matrix)
* - `projection_matrix` (P, 3x4 double matrix)
* - `local_transform` (camera pose w.r.t robot frame)
*
* On success, the internal matrices and settings of the camera model are updated. If the model
* is valid for rectification, the rectification map is initialized.
*
* @param filePath Absolute or relative path to the YAML file.
* @param initRectificationMaps Set to false to skip building the (potentially large) rectification
* maps when rectification won't be used (saves time and memory).
* @return True if the file was successfully loaded and parsed, false otherwise.
*
* @warning Logs warnings if any fields are missing. If file does not exist or parsing fails, returns false.
*
* @see initRectificationMap()
*/
bool load(const std::string & filePath, bool initRectificationMaps = true);
/**
* @brief Loads the camera model by constructing a file path from a directory and camera name.
*
* This is a convenience wrapper around `load(filePath)` that constructs the file path as:
* `directory + "/" + cameraName + ".yaml"`.
*
* @param directory Path to the folder containing the camera YAML file.
* @param cameraName Base name of the camera file (without extension).
* @param initRectificationMaps Set to false to skip building the (potentially large) rectification
* maps when rectification won't be used (saves time and memory).
* @return True if loading from the constructed path succeeds, false otherwise.
*/
bool load(const std::string & directory, const std::string & cameraName, bool initRectificationMaps = true);
/**
* @brief Saves the camera model parameters to a YAML calibration file in ROS format.
*
* The file will include the following fields if they are not empty:
* - `camera_name`
* - `image_width`
* - `image_height`
* - `camera_matrix` (K)
* - `distortion_coefficients` (D)
* - `distortion_model` (auto-detected based on number of distortion coefficients)
* - `rectification_matrix` (R)
* - `projection_matrix` (P)
* - `local_transform` (camera pose w.r.t robot frame)
*
* If the distortion matrix contains 6 coefficients (used for fisheye), it is converted
* to a standard 4-coefficient format for ROS compatibility.
*
* @param directory Path to the folder where the YAML file will be saved.
* @return True if saving was successful, false otherwise.
*
* @note If `name_` is empty, "camera.yaml" is used as the default filename.
* @warning Returns false and logs an error if none of the matrices are set.
*/
bool save(const std::string & directory) const;
/**
* @brief Serializes the camera model to a binary format.
*
* The serialization includes the camera intrinsics (`K_`, `D_`), rectification matrix (`R_`),
* projection matrix (`P_`), image size, and the local transform. The format is compact and suitable
* for file storage or transmission over a network.
*
* Data layout:
* - Header (11 integers):
* - [0-2] RTAB-Map version (major, minor, patch)
* - [3] Camera type (0 = mono, 1=stereo)
* - [4-5] Image width, height
* - [6-9] Element counts for K, D, R, P matrices
* - [10] Size of localTransform (0 if null)
* - Data section (in order): raw memory blocks for K, D, R, P (`double` values), followed by `float` values for localTransform
*
* @return A byte vector containing the serialized data. The format is compatible with `deserialize()`.
*
* @note This is a custom binary format, not meant to be human-readable.
* @see deserialize(), StereoCameraModel
*/
std::vector<unsigned char> serialize() const;
/**
* @brief Deserializes a camera model from a byte vector.
*
* This is a convenience wrapper around `deserialize(const unsigned char*, unsigned int)`
* that takes a `std::vector<unsigned char>` instead of a raw buffer.
*
* @param data Byte vector containing data serialized by `serialize()`.
* @return The number of bytes successfully read and parsed. Returns 0 on failure.
*
* @see serialize(), deserialize(const unsigned char*, unsigned int)
*/
unsigned int deserialize(const std::vector<unsigned char>& data);
/**
* @brief Deserializes a camera model from a raw byte buffer.
*
* Reads the camera intrinsics, distortion, rectification, projection matrices, image size,
* and local transform from a serialized binary format previously created with `serialize()`.
*
* @param data Pointer to the binary data buffer.
* @param dataSize Size of the data buffer in bytes.
* @return The number of bytes successfully read. Returns 0 on error or if the format is invalid.
*
* @warning If the buffer format does not match the expected layout or version, an error is logged
* and the camera model remains in a default-initialized state.
*
* @note Assumes little-endian architecture and strict size/type matching. The serialized format
* must be created by `CameraModel::serialize()`. Non-mono camera types are not supported.
* See `StereoCameraModel` to serialize/deserialize stereo models.
*
* @see serialize()
*/
unsigned int deserialize(const unsigned char * data, unsigned int dataSize);
/**
* @brief Returns a new camera model with all intrinsic parameters scaled by a given factor.
*
* This method scales the camera's intrinsic matrix (`K_`) and projection matrix (`P_`), as well as
* the image size, by the given `scale` factor. The distortion and rectification matrices are left unchanged.
*
* Only valid camera models (i.e., those for which `isValidForProjection()` returns true) are scaled.
* If the model is invalid, a warning is issued and the original model is returned unchanged.
*
* @param scale Scaling factor (> 0). For example, use 0.5 to downscale or 2.0 to upscale.
* @return A scaled copy of the camera model with updated intrinsics and image size.
*
* @warning If the camera model is not valid for projection, the scale operation is ignored.
*/
CameraModel scaled(double scale) const;
/**
* @brief Returns a new camera model adjusted for a given region of interest (ROI).
*
* This method shifts the principal point (`cx`, `cy`) in the intrinsic matrix (`K_`) and projection matrix (`P_`)
* by subtracting the ROI’s top-left `(x, y)` offset. The image size is also set to the ROI size.
*
* Only valid camera models (i.e., those for which `isValidForProjection()` returns true) can be adjusted.
* If the model is invalid, a warning is issued and the original model is returned unchanged.
*
* @param roi Region of interest defined as a rectangle (typically a subwindow of the full image).
* @return A new camera model adapted to the ROI with adjusted intrinsics and image size.
*
* @warning If the camera model is not valid for projection, the ROI operation is ignored.
*/
CameraModel roi(const cv::Rect & roi) const;
/**
* @brief Rectifies a raw image using the precomputed rectification maps.
*
* This function applies geometric correction (rectification) to an image using the camera model's
* `mapX_` and `mapY_` rectification maps. It is typically used to correct lens distortion in images
* based on the calibration parameters.
*
* @param raw Input raw image (e.g., from camera). Must be a valid `cv::Mat`.
* @param interpolation Interpolation method to use. Typically `cv::INTER_LINEAR` or `cv::INTER_NEAREST`.
*
* @return Rectified image. If the rectification maps are not initialized, the function logs an error
* and returns a clone of the original image.
*
* @pre `mapX_` and `mapY_` must be initialized using `initRectificationMap()`.
*
* @note Works for color and grayscale images of any valid type.
*
* @see initRectificationMap(), rectifyDepth()
*/
cv::Mat rectifyImage(const cv::Mat & raw, int interpolation = cv::INTER_LINEAR) const;
/**
* @brief Rectifies a raw depth image using the precomputed rectification maps.
*
* This function applies geometric correction (rectification) to a 16-bit unsigned depth image.
* It performs a pixel-by-pixel bilinear interpolation, only if all neighboring pixels have valid
* (non-zero) depth values, and the variation among them is within 1% of their average.
*
* The method is optimized to avoid introducing noise in regions of high depth variance.
*
* @param raw Input raw depth image (`CV_16UC1`). Must contain 16-bit unsigned depth values.
*
* @return Rectified depth image. If the rectification maps are not initialized or the input
* image is not of type `CV_16UC1`, the function logs an error and returns a clone of the input.
*
* @pre Input image must be of type `CV_16UC1`. `mapX_` and `mapY_` must be initialized.
*
* @note Inspired by the Kinect2 CPU depth registration implementation from:
* https://github.com/code-iai/iai_kinect2
*
* @warning Invalid or noisy regions are skipped in interpolation to maintain depth consistency.
*
* @see initRectificationMap(), rectifyImage()
*/
cv::Mat rectifyDepth(const cv::Mat & raw) const;
/**
* @brief Projects a 2D pixel and depth value into a 3D point in the camera coordinate frame (/camera_link).
*
* This function uses the camera's intrinsic parameters to compute the 3D point corresponding to the given
* 2D image coordinates and depth value.
*
* @param u Horizontal image coordinate (in pixels).
* @param v Vertical image coordinate (in pixels).
* @param depth Depth value at (u, v) in meters.
* @param[out] x Output X coordinate in 3D space.
* @param[out] y Output Y coordinate in 3D space.
* @param[out] z Output Z coordinate in 3D space (equals `depth`).
*
* @note If `depth <= 0`, the output (x, y, z) will be set to `NaN`.
*
* @see reproject()
*/
void project(float u, float v, float depth, float & x, float & y, float & z) const;
/**
* @brief Reprojects a 3D point in the camera frame (/camera_link) into 2D image coordinates (floating-point).
*
* This function computes the image plane coordinates for a given 3D point using the camera's
* intrinsic parameters.
*
* @param x X coordinate in camera space.
* @param y Y coordinate in camera space.
* @param z Z coordinate in camera space (must be non-zero).
* @param[out] u Output horizontal image coordinate (float).
* @param[out] v Output vertical image coordinate (float).
*
* @pre `z != 0`
*
* @see project(), reproject(int&, int&)
*/
void reproject(float x, float y, float z, float & u, float & v) const;
/**
* @brief Reprojects a 3D point in the camera frame (/camera_link) into 2D image coordinates (rounded to int).
*
* This version of `reproject()` returns integer pixel indices, computed from the 3D position.
*
* @param x X coordinate in camera space.
* @param y Y coordinate in camera space.
* @param z Z coordinate in camera space (must be non-zero).
* @param[out] u Output horizontal image coordinate (integer pixel).
* @param[out] v Output vertical image coordinate (integer pixel).
*
* @pre `z != 0`
*
* @see project(), reproject(float&, float&)
*/
void reproject(float x, float y, float z, int & u, int & v) const;
/**
* @brief Checks if a given pixel coordinate lies within the image bounds.
*
* @param u Horizontal image coordinate (in pixels).
* @param v Vertical image coordinate (in pixels).
* @return `true` if the pixel is within the image dimensions, `false` otherwise.
*
* @note Inclusive lower bound, exclusive upper bound: `[0, width)`, `[0, height)`
*/
bool inFrame(int u, int v) const;
private:
std::string name_; ///< Camera name.
cv::Size imageSize_; ///< Image size.
cv::Mat K_; ///< Intrinsic matrix.
cv::Mat D_; ///< Distortion coefficients.
cv::Mat R_; ///< Rectification matrix.
cv::Mat P_; ///< Projection matrix.
cv::Mat mapX_; ///< Rectification map X.
cv::Mat mapY_; ///< Rectification map Y.
Transform localTransform_; ///< Transform from camera to base link.
};
/**
* @brief Stream operator for printing a camera model to an output stream.
*
* This function outputs the name, image size, and camera matrices (K, D, R, P)
* along with the local transformation.
*
* Example output:
* ```
* Name: camera1
* Size: 640x480
* K= [fx, 0, cx;
* 0, fy, cy;
* 0, 0, 1]
* D= [...]
* R= [...]
* P= [...]
* LocalTransform= [...]
* ```
*
* @param os Output stream.
* @param model Camera model to print.
* @return The modified output stream.
*/
RTABMAP_CORE_EXPORT std::ostream& operator<<(std::ostream& os, const CameraModel& model);
} /* namespace rtabmap */
#endif /* CAMERAMODEL_H_ */