The car is considered a rig with several rigidly-attached sensors. The dimensions of the car and the positions and orientations of these sensors relative to the car are important to NVIDIA® DriveWorks accuracy. The particular properties of a rig, which are represented by a rig configuration, are measured and estimated by a calibration process.
The rig configuration module allows reading and enumerating of these pre-calibrated properties. The module obtains the rig configuration from a rig configuration file generated by the DriveWorks calibration tool. For more information, see DriveWorks Calibration Tool Application Note in the doc/pdf/calibration folder of this release. Rig configuration files can also be serialized by the module.
An example of the JSON structure of a rig configuration file is given below.
A rig configuration object (dwRigHandle_t) can be initialized either from a file or directly from string (see dwRig_initializeFromFile()). Once the rig configuration is loaded successfully, individual properties of the rig and it's sensors can be queried. Vehicle properties are represented by the dwVehicle struct (see dwRig_getVehicle()). Generic sensor properties can be obtained through the function dwRig_getSensorXXX. Some sensors may have more specific properties, which can be obtained using dedicated functions, like the calibrated camera model (for example see dwRig_getPinholeCameraConfig()).
A number of modules use a camera-only rig that contains only rigidly-attached calibrated cameras (dwCameraModelHandle_t). This camera rig is a subset of the generic rig. It can be initialized independently or directly from a generic rig configuration (see dwRig_initializeFromFile()).
The supported calibrated camera models are pinhole, OCam, and FTheta. Their calibration parameters are specified with the dwPinholeCameraConfig, and dwFThetaCameraConfig structs, respectively. For each of these camera models, DriveWorks provides generic dwCameraModel_pixel2Ray() and dwCameraModel_ray2Pixel() functions to transform a point from image space to camera space and vice versa (see Intrinsic Camera Models for details).
