IEEE Robotics & Automation Magazine - June 2020 - 158

Once the image is uploaded, the processing can be done as
usual, and the result can be downloaded and converted into
an ROS message.
Conclusions
CUDA for OpenCV is an easy solution for accelerating vision
applications in robotics on systems equipped with a CUDAenabled GPU. The migration of the code from CPU-based to
GPU-based is simple and relatively straightforward, even trivial in some cases. The speedup that can be achieved is systemand problem-dependent.
For simple vision algorithms, modern CPUs
An advantage of an open
can be faster; for complex
problems involving a
framework such as OpenCL
sequence of operations
on the image, parallelizaover CUDA is that it is
tion in the GPU leads to
supported by both AMD and better performance; and
for deep-learning applications, the improvement
Nvidia cards.
is significant.
We provided some examples with well-known
algorithms that are widely used by the robotics community,
with the aim of encouraging researchers to improve the
throughput of their systems by squeezing all the computing
power out of their hardware. CUDA and other computing
frameworks (DirectCompute [15] and OpenCL [16]) have
become programming standards for parallel computing, and
their inclusion in popular libraries like OpenCV is an opportunity for developers to benefit from parallelization without a
significant investment in learning specific parallel programming techniques.
An advantage of an open framework such as OpenCL over
CUDA is that it is supported by both AMD and Nvidia cards.
The interested reader can refer to [17] for details about using
OpenCL in OpenCV.
References
[1] D. Luebke, "CUDA: Scalable parallel programming for high-performance scientific computing," in Proc. 2008 5th IEEE Int. Symp. Biomedical Imaging: From Nano to Macro, pp. 836-838. doi: 10.1109/
ISBI.2008.4541126.
[2] K. Pulli, A. Baksheev, K. Kornyakov, and V. Eruhimov, "Real-time
computer vision with OpenCV," Commun. ACM, vol. 55, no. 6, pp.
61-69, June 2012. doi: 10.1145/2184319.2184337.
[3] P. Michel, J. Chestnutt, S. Kagami, K. Nishiwaki, J. Kuffner, and T.
Kanade, "GPU-accelerated real-time 3D tracking for humanoid locomotion and stair climbing," in Proc. 2007 IEEE/RSJ Int. Conf. Intelligent Robots and Systems, pp. 463-469. doi: 10.1109/IROS.2007.4399104.
[4] T. Xu, T. Pototschnig, K. Kuhnlenz, and M. Buss, "A high-speed
multi-GPU implementation of bottom-up attention using CUDA," in

158

*

IEEE ROBOTICS & AUTOMATION MAGAZINE

*

JUNE 2020

Proc. 2009 IEEE Int. Conf. Robotics and Automation, pp. 41-47. doi:
10.1109/ROBOT.2009.5152357.
[5] J. Kim, E. Park, X. Cui, H. Kim, and W. A. Gruver, "A fast feature
extraction in object recognition using parallel processing on CPU and
GPU," in Proc. 2009 IEEE Int. Conf. Systems, Man and Cybernetics, pp.
3842-3847. doi: 10.1109/ICSMC.2009.5346612.
[6] N. Dalmedico, M. A. S. Teixeira, H. S. Barbosa, A. S. de Oliveira,
L. V. R. de Arruda, and, and F. Neves Jr, "GPU and ROS: The use of
general parallel processing architecture for robot perception," in
Robot Operating System (ROS) (Studies in Computational Intelligence), A. Koubaa Ed. pp. 407-448. Cham, Switzerland: Springer
-Verlag, 2018.
[7] I. Culjak, D. Abram, T. Pribanic, H. Dzapo and M. Cifrek, "A brief
introduction to OpenCV," in Proc. 35th Int. Conv. MIPRO, 2012, pp.
1725-1730.
[8] J. Canny, "A computational approach to edge detection," IEEE Trans.
Pattern Anal. Mach. Intell., vol. 8, no. 6, pp. 679-698, 1986. doi: 10.1109/
TPAMI.1986.4767851.
[9] R. Mur-Artal and J. D. Tardós, "ORB-SLAM2: An open-source
SLAM system for monocular, stereo, and RGB-D cameras," in IEEE
Trans. Robot., vol. 33, no. 5, pp. 1255-1262, Oct. 2017. doi: 10.1109/
TRO.2017.2705103.
[10] G. Farneback, "Very high accuracy velocity estimation using orientation tensors, parametric motion, and simultaneous segmentation
of the motion field," in Proc. 2011 IEEE Int. Conf. Computer Vision, pp.
171-177. doi: 10.1109/ICCV.2001.937514.
[11] Y. Jia et al., "Caffe: Convolutional architecture for fast feature
embedding," in Proc. 22nd ACM Int. Conf. Multimedia, pp. 675-678.
2014, doi: 10.1145/2647868.2654889.
[12] M. Abadi et al., "Tensorflow: A system for large-scale machine
learning," in Proc. 12th USENIX Symp. Operating Systems Design and
Implementation (OSDI), pp. 265-283. 2016.
[13] J. Redmon and A. Farhadi, Yolov3: An incremental improvement.
2018. [Online]. Available: https//arXiv:1804.02767
[14] M. Quigley et al., "ROS: An open-source robot operating system,"
in Proc. ICR A Workshop on Open Source Software, 2009, vol. 3,
no. 3.2, p. 5.
[15] T. Ni, "Direct Compute: Bring GPU computing to the mainstream," in Proc. GPU Technology Conf., 2009, p. 23.
[16] P. Du, R. Weber, P. Luszczek, S. Tomov, G. Peterson, and J. Dongarra, "From CUDA to OpenCL: Towards a performance-portable solution for multi-platform GPU programming," Parallel Comput., vol. 38,
no. 8, pp. 391-407, 2012. doi: 10.1016/j.parco.2011.10.002.
[17] H. Gasparakis, "Heterogeneous compute in computer vision:
OpenCL in OpenCV," in Proc. SPIE 9029, Visual Information Processing and Communication V. 2014. vol. 9029, pp. 104-111. doi:
10.1117/12.2054961.

Enric Cervera, Department of Computer Science and Engineering, Jaume-I University, Castelló de la Plana, Castellón,
Spain. Email: ecervera@uji.es.
IEEE Robotics & Automation Magazine - June 2020

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - June 2020
