EYE on NPI – Intel RealSense LIDAR camera #EyeOnNPI #LIDAR #DigiKey @DigiKey @Intel @Adafruit: A New video by Adafruit Industries

Almost 10 years ago, Adafruit sponsored a Kinect-hacking bounty (https://ift.tt/1F7up0t) to take an off-the-shelf video game accessory and turn it into a tool for anyone to use. The open source cross-platform drivers spawned hundreds of awesome projects, from the scientific and artistic, to the humanist and just plain silly (https://ift.tt/2ZMyGuL). With the Kinect discontinued (https://ift.tt/1j7piVI), it’s been a challenge to find a low cost alternative with an easy API. Until now! This little sensor is smaller than a tuna can and is a complete system with plug-and-play USB type C port for power and data. There’s even an BMI085 IMU inside! (https://ift.tt/35Mg8im)

Intel RealSense’s L515 is a type of solid state LiDAR technology-based depth camera. It uses LiDAR technology to scan a full scene allowing it to provide a point cloud with up to 23 million points of accurate depth data per second. It delivers consistent high depth accuracy throughout the supported range of the camera starting at 25 cm up to 9 meters. The accuracy of the L515 within its operational range is less than 20 mm Z error at maximum range. With an internal vision processor, L515 achieves motion blur artifact reduction by an exposure time of less than 100 ns (nanoseconds), ideal for capturing rapidly moving objects with minimal motion blur. Additionally, the 4 ms photon to depth latency allows real-time applications such as autonomous navigation. An RGB camera, gyroscope, and accelerometer round out the capabilities of the device for high-quality performance and use case flexibility.

Smaller than a tennis ball, the Intel RealSense LiDAR Camera L515 has a diameter of 61 mm and is 26 mm in height. Weighing in at around 100 g (3.5 oz), it is designed to be easily situated on any prototype or attached to a tablet or phone for use in handheld room scanning or volumetric measurement applications.

The L515 uses the same open source Intel RealSense SDK 2.0 as all other current-generation Intel RealSense Depth Cameras, which is platform independent, supporting Windows, Linux®, Android™, and macOS. This includes wrappers for many common platforms, languages and engines, including Python, ROS, C/C++, C#, Unity, Unreal, OpenNI, and NodeJS, with more being added constantly.

Features
Depth FOV (field of view): (H × V) 70°±2° x 55°±2°
Depth resolution: Up to 1024 x 768 depth resolution (XGA resolution)
Depth frame rate: 30 frames per second
RGB frame rate: 30 frames per second
RGB resolution: Up to 1920 x 1080 RGB resolution
RGB FOV: (H x V) 70° ±3° x 43° ±2°
Use environment: Indoor
Inertial measurement unit: Bosch BMI085
Range: From 25 cm to 9 m range
Connection type: USB-C® 3.1 gen 1 connection
Dimensions: 61 mm diameter x 26 mm height
Mounting: Two M3 x 0.5 mounting points and one 1/4 20 UNC thread mounting point

It shows up in a few ways when you plug it into your computer. For integration into projects that use OpenCV or other camera-based inputs, you will see a dual UVC camera one called ‘Depth’ which is the LIDAR and one called ‘RGB’ which is a regular camera that you can overlay over the depth data. The Cameras are really high res, as well – 1024×768 depth resolution and 1920 x 1080 RGB resolution. Compare that to the Kinect’s 640×480! (https://ift.tt/1j7piVI)

We like that this sensor has industrial-quality sensing with engineer-and-maker-friendly development. No NDAs for datasheets, no signups for code – everything needed to get set up was shockingly convenient. (https://ift.tt/33FluZI) There’s even a developer center (https://ift.tt/3ceBCp5) with firmware upgrades, documentation, code samples, and a community group. There’s a lot of support, so we think this is a great option for low cost depth sensing – a really fabulous NPI!