deviceCalib->getCameraIntrinsics still necessary with updates?

AdamPolak

I am updating an old project and I wanted to find the spatial x,y,z coordinates of an object on host using detections and a sent depth map.

I currently have it as:

deviceCalib = std::make_shared[dai::CalibrationHandler](dai::CalibrationHandler)(device->readCalibration());
cameraIntrinsics = deviceCalib->getCameraIntrinsics(dai::CameraBoardSocket::RGB, 1280, 720);
auto lensPos = deviceCalib->getLensPosition(dai::CameraBoardSocket::RGB);
cameraHFOVInRadians = ((deviceCalib->getFov(dai::CameraBoardSocket::RGB) * pi) / 180.0) * (3840.0/4056.0); // Must scale for cropping: https://discordapp.com/channels/790680891252932659/924798503270625290/936746213691228260

Is this still the right process?

jakaskerl

Hi AdamPolak
Check https://github.com/luxonis/depthai-experiments/tree/master/gen2-calc-spatials-on-host (main.py and calc.py) to see how coordinates are being calculated from the depth map. It should be the same for C++.

Thanks,
Jaka

AdamPolak

Much appreciated, looks like there is a new way to get these values:

# Required information for calculating spatial coordinates on the host HFOV = np.deg2rad(self.calibData.getFov(dai.CameraBoardSocket(depthData.getInstanceNum())))