Calibrateing Camera to IMU in Matlab

I need to calibrate my flire blackflyS camera to my IMU but I am having trouble doing so using the estimateextrinsics. I know this is done through kalibr but due to the lens we are using the kalibr system is unable to calibrate our system. Does anyone know how to do this on matlab? I saw on a previous post that they imu adn camera can be taken in as as arguments for the estimateextrinsics function but I dont know how to do so and could use advice

Answer 1

[Venkat Siddarth]

Hi John,

A new function has been introduced to achieve this estimation in MATLAB R2024a. It is called estimateCameraIMUTransform. It is used to estimate transformation from camera to IMU using calibration data.

To learn more about the function, check out the documentation: https://in.mathworks.com/help/nav/ref/estimatecameraimutransform.html

Regards

Venkat Siddarth V.

Answer 2

[Akshai Manchana]

Calibrating a camera-IMU extrinsic is usually an iterative process.

• During data collection we need to excite all sensor axes enough to gather good accelometer and gyroscope measurement. Otherwise noise present in the IMU data will dominate and results in lower calibration accuracy. Follow a few recommondations provided in Camera-IMU Extrinsic Calibration Example
• The IMU and camera time stamps must be sampled from a single clock and be accurate for greater accuracy.
• Camera intrinsic paramater, IMU noise parameter estimation must also be accurate.

Following the recommondations from the example can help you accurately estimate the transformation between camera and IMU accurately. Looking at the actual calibration data can help us in coming up with better suggestions and help you go through the calibration using the utility function estimateCameraIMUTransform from Navigation Toolbox.