Inquiry Regarding Suitable RealSense Camera for Industrial Wood Dimension Measurement Project
Dear RealSense Technical Support Team,
I hope this message finds you well.
I am currently working on an industrial project that involves measuring the physical dimensions of wooden components within a manufacturing environment. The primary objective is to accurately measure the length, width, and possibly height of wooden pieces using a camera-based vision system.
For this application, the system will be deployed in an industrial setting, and accuracy and reliability are important considerations. I am looking to select an appropriate RealSense depth camera that can provide precise dimensional measurements and operate consistently in such an environment.
Additionally, the camera should be compatible with embedded processing platforms such as Raspberry Pi or NVIDIA Jetson Nano, as the image processing and measurement algorithms will be executed on one of these platforms.
I would appreciate your guidance in selecting the most suitable RealSense camera model based on the following requirements:
-
Industrial indoor environment usage
-
Accurate measurement of wooden object dimensions
-
Compatibility with Raspberry Pi or Jetson Nano
-
Ease of integration and configuration
Please let me know if you require any further technical details regarding the project to recommend the most appropriate camera solution.
Thank you for your time and support. I look forward to your response.
Kind regards,
Aakash Parmar
Embedded Hardware Engineer
-
Hi Aakash Parmar If the camera will be positioned more than half a meter from the observed wood surface and will be operating in an industrial environment then the D456 RealSense model will likely meet your needs.
It has an IP65 rated casing that protects against entry by dust or projected water (e.g rain), a wide field of view, has very good depth accuracy over distance (around 2x that of the D435 type models) and is equipped with an infrared dot pattern projector component that aids analysis of smooth plain surfaces for depth information.
https://store.realsenseai.com/buy-intel-realsense-depth-camera-d456.html
In regard to choice of computing hardware to use with the camera, I would recommend Jetson Nano (ideally the newer Orin Nano). It is possible to use the camera with Raspberry Pi (ideally Pi 5) but whilst Jetson is officially supported by RealSense, Pi is not.
If your system will be using an RGB colour image to calculate dimensions instead of depth information, the new D555 model would be worth considering, It has the same benefits as D456 plus a new-generation data processing circuit board that enhances RGB with Geometric Distortion Correction (GDC) and Temporal Noise Reduction (TNR). It also has built-in Power Over Ethernet (PoE) ethernet network cabling support, enabling the camera to be used up to 100 meters from the computing device over ethernet instead of on a short USB cable.
-
Hi MartyX Grover,
Thank you for your feedback.
I would like to understand the accuracy levels of the Intel RealSense D435 and D455 models, as our application involves measuring the cant of wood in an industrial environment.
Could you please advise whether these cameras are suitable for this use case and whether they can provide reliable and accurate measurements for this type of data?
Looking forward to your guidance. -
-
D455 has the same performance as D456 but without the dust / water resistant casing. So if the environment that the wood is being scanned in does not have dust / sawdust then D455 would be fine.
D435 has lower accuracy over distance than D455 / D456. Error starts at around zero at the front glass of the camera and increases linearly as distance of an observed object from the camera increases. So if the wood is being observed at relativity close range (less than a meter from the camera) then the accuracy difference between D435 and D455 / D456 will not be significant.
D435 has a slower shutter on its RGB sensor which can result in lag / blur on the RGB image when the camera is moving quickly or fast motion is being observed. But if the wood is stationary, or you are not using RGB data, then D435 would also be an appropriate choice for your application.
My research indicates that a cant is a lumber log that has been sawn on 4 sides into a square shape. In that regard, it sounds similar to measuring the three-dimensional volume of a box, which RealSense cameras can certainly do well.
The official free RealSense Viewer tool has an easy to use measuring mode where you can select 2 or more points on a depth image and be provided with the distance between those points.
If you are scanning a long sawn lumber log and have to have a longer distance from the camera to fit the entire log into the camera's field of view then the wider view and higher accuracy over distance of D455 / D456 will be recommendable.
Please sign in to leave a comment.
Comments
4 comments