O. Zendel, C. Sulzbachner, and W. Kubinger (Austria)
Modelling, Control, High Dynamic RangeCameras, Computer Vision, Robotics
We present an intelligent controller for high dynamic range (HDR) cameras using a model based approach. HDR cameras capture all scenes with high contrast even under difficult illumination conditions. For this purpose special camera parameters have to be set according to the scenery to maximize the contrast of the output image. The most essential part of the controller is a correct camera model. This is achieved by taking test images using a dedicated measurement assembly. The images are transformed into data point clouds which are converted into the camera model. The calculated camera model can be used to predict the effects of camera parameter changes on the output images and allows the calculation of fitting camera parameters for the current scenery. This way the controller achieves an adaptation of the sensor to the environment. A prototype implementation of our intelligent controller has been realized using an FPGA development board. Experimental results show that the controller is suitable for real-time vision sensors for outdoor applications.
Important Links:
Go Back