By attaching a Torobo Eye on the tip of a robot arm or the head of a dual-arm robot, a compact autonomous robot can be constructed for the applications such as bin picking and assembly. Especially, collaborative robots (which do not require safety fences) and mobile manipulators mounting a Torobo Eye can do those tasks anywhere in the field because any special devices don't need to be installed in the environment. This is a big advantage compared to a stationary-type 3D camera.
Since Torobo Eye uses a single camera to acquire both depth and color images, there is no misalignment between the two images, and highly accurate 3D color information can be obtained. Therefore, by combining with recognition technology that uses machine learning and/or image processing, Torobo Eye can be used for the applications such as visual inspection, shape measurement, and quality check. Its small size allows for a high degree of flexibility in installation, making it possible to build space-saving inspection and production lines.
The High Dynamic Range (HDR) measurement allows you to capture images at multiple exposure times and integrate pixels with good depth quality from each exposure-time image. This makes it possible to measure metallic luster and black workpieces that are prone to missing depth in a single exposure.
ROS package has been released for Torobo Eye. By using this, it is easy to use Torobo Eye with ROS.
The Torobo Eye SDK (Python, C ++) includes installation scripts for the OpenCV and Open3D libraries, and you can immediately implement recognition processing using these libraries.
SL40 | SL80 | description | |
Measurement method | Active Stereo (Structured Light) | Algorithm based on a phase shifting method | |
Base color of projection pattern | White, Red, Green, Blue | Can be changed according to the color of target object | |
Resolution [pixel] | 1440 x 1080 | ||
Depth error [mm] | 0.06 | 0.10 | SL40: Standard deviation of measurements at 450mm distance with high resolution mode SL80: Standard deviation of measurements at 700mm distance with high resolution mode |
Measurement time [msec] | 150 | In the case of one shot measurement with high speed mode | |
Measurement range (XY) [mm] | 375 x 210 | 409 x 306 | SL40: Valid XY range at 450mm distance SL80: Valid XY range at 700mm distance |
Measurement range (Z) [mm] | 300 - 600 | 400 - 1200 (500 - 1000 recommended) |
Recommended range for depth direction |
Supported OS | Ubuntu 18.04 LTS, Ubuntu 20.04 LTS, Windows 10/11 | OS on your PC that can connect to Torobo Eye | |
Software | UI software, SDK (API library), ROS package | API libraries support Python and C++. | |
Communication I/F | Ethernet | Gigabit Ethernet or better is necessary. | |
Power supply | 24V DC | ||
Power consumption (continuous) [W] | 48 | Total power consumption of camera head and controller on average while projecting and imaging continuously | |
Power consumption (idle) [W] | 12 | Total power consumption of camera head and controller in the idle state (no measurement) | |
Dimension [mm] | Camera Head: W92 x H75.5 x D95 Camera Controller: W140 x H50 x D140 |
Camera Head: W140 x H47 x D97 Camera Controller: W140 x H50 x D140 |
The size of the connector and cable is not considered. |
Weight [g] | Camera Head: 460 Camera Controller: 854 |
Camera Head: 500 Camera Controller: 854 |
You can acquire the camera images by connecting a Torobo Eye to your own PC via Ethernet. The power supply unit (e.g. AC-DC power adapter) is not provided, so please prepare a power source such as a switching power supply and battery that generates 24V DC by yourself. It is also possible to input 24V directly to the camera head from a separate power supply.
For industrial applications, Torobo Eye can be easily connected from your own software by using the provided SDK. Since Torobo Eye only outputs 3D data, you need to implement your own algorithms such as object recognition on your PC if necessary. The ROS package is also provided, so if you use a ROS-compatible robot, you can build a robot system more easily.
Torobo Eye is used for Yamaha Motor's bolt picking system, which was exhibited at the 2022 International Robot Exhibition (iRex 2022).
The following points were highly evaluated:
- Sufficient detection accuracy for bolt picking (M3 size and up)
- Easy for operators to handle it due to its non-laser light source
- High cost performance