OpenCV is the most widely used library for image processing. Some of the reasons for its worldwide popularity is because of its ease of use, well-written documentation, easy portability and most importantly a large and active community to support its development. Earlier, we had posted a blog/article to enable our customers to access e-con’s MIPI cameras for the Jetson development kits using OpenCV.
Almost all our cameras can be accessed natively in OpenCV, but the performance is not always the best when used as per OpenCV’s official documentation. We have been asked the same questions by our users multiple times: “How do I access the camera in OpenCV?”, “I can’t use the camera’s highest frame rates in OpenCV. What am I missing?”, “Is it possible to get 4K@30 fps from your camera using OpenCV framework?“. This article is an attempt to help our users get the maximum performance from our cameras when accessed in OpenCV. Check it out here: Accessing cameras in OpenCV with high performance
Although the targeted platform in the article is the Jetson TX1/TX2 (aarch64), many of the options used are the same for other platforms as well. Also, we used the e-CAM130_CUTX1 for evaluation but any other V4L2 camera (Yes, USB cameras too!) can be accessed using the methods described in the article.
Dilip Kumar is a computer vision solutions architect having more than 8 years of experience in camera solutions development & edge computing. He has spearheaded research & development of computer vision & AI products for the currently nascent edge AI industry. He has been at the forefront of building multiple vision based products using embedded SoCs for industrial use cases such as Autonomous Mobile Robots, AI based video analytics systems, Drone based inspection & surveillance systems.