To access the content on this page, make sure JavaScript is enabled in your browser. Depth cameras play a crucial role in providing robots, drones, and autonomous systems with a comprehensive understanding of their surroundings. These sensors enable the measurement of distances to nearby objects and the creation of 3D maps, empowering machines to navigate, avoid obstacles, and interact effectively with their environment.
Deploying a depth camera, like the latest RealSense D585 model, is just the initial step. Engineers must then develop algorithms and control systems that leverage the data captured by these cameras to power various functionalities. The D585 stands out for simplifying computer vision applications by offering superior image quality and conducting significant image processing on the device itself, allowing developers to focus on innovation rather than intricate technicalities.
At the heart of the D585 is the RealSense V5 system-on-chip, an all-in-one package comprising a depth engine, image signal processor, DSP, AI hardware acceleration, and a quad-core Arm processor. Unlike conventional cameras that rely on host computers for processing raw data, the D585 can execute multiple tasks internally. Described as an AI-native, software-defined vision platform, this device can receive new capabilities through SDK updates post-deployment.
Equipped with dual infrared projectors, high-resolution sensors, and global shutter imaging, the stereo depth camera offers robust indoor and outdoor performance. Its wide 120° by 100° field of view provides robots with a comprehensive perspective of their surroundings, with output resolutions of up to 1280 × 960 pixels. Capable of running at 60 frames per second in full resolution and registering depth at distances as close as 10 cm, the D585 fits various applications, including robotic manipulation, inspection systems, and collaborative robot arms operating near objects.
The D585 boasts an impressive operating range of over 20 meters, making it suitable for supporting mobile robots in expansive warehouses and industrial settings. It introduces a Dual RGB mode for synchronizing RGB and depth streams seamlessly while handling stream alignment within the camera itself. Additionally, the Depth Compression feature reduces bandwidth requirements by approximately 75%, particularly useful for transmitting high-frame-rate depth data across constrained links.
RealSense plans to enhance the D585’s capabilities further through updates in its SDK ecosystem. The camera will initially offer person detection in beta, with upcoming features such as visual-inertial odometry, occupancy grid generation, auto-calibration, face detection, and improved depth processing. The expected shipping date for the D585 is the first quarter of 2027, with pricing details yet to be announced.
About the Author
