Our LiDAR sensor streams fascinating 3D point clouds that allow a host of innovative applications. However, for some applications, 3D point cloud processing can be too expensive. What if we could represent our data as 2D image planes and use less expensive image processing algorithms on them? How do convolutional neural networks perform on this type of data? Join our team of engineers and prove that you are able to push the boundaries of what is possible!
• Implement a FPGA module to stream depth images
• Record data for algorithm development • Evaluate algorithms for denoising, undistortion, etc.
• Evaluate approaches for machine learning, e.g. CNN
• Support marketing to prepare impressive demos
• Experience with embedded hardware, especially FPGA
• Programming skills (VHDL, Python, C/C++)
• Experience with image processing; machine learning frameworks is a plus
• First experience with (embedded) Linux