This page provides detailed information about the built-in Object Detection Function.
This feature is available on ugo Pro (manufactured after March 2025) and ugo mini (G1) running ugo OS v4.2.0 or later.
*If you wish to use the Object Detection Function, we will need to remotely apply the detection AI model to your robot, in addition to this update.
If you wish to use the Object Detection Function on an applicable model, please contact us .
Table of Contents
・What is the Object Detection Function?
・Details of the "Object Detection" command
・Parameter setting example for "Object Detection" command
・Specifying the range of "Object Detection"
・Example of creating a Flow using the Object Detection Function
What is the Object Detection Function?
The Object Detection Function is a function that detects the presence of people in the surrounding area in real time based on information obtained from ugo's sensors such as cameras and LiDAR.
When a person is detected within the configurable detection range, this function is automatically activated and proceeds to the next command according to the pre-set flow.
◆ Changes in object detection
| Previous "Person Detection (Camera)" command | New "Object Detection" command | |
| Compatible models | ugo Pro/Ex (manufactured up to February 2025) | ugo Pro/Ex (manufactured after March 2025) mini (G1) |
| Analysis Speed | Every second | Every 0.1 seconds |
| Range Specification | Width and depth specification | Distance only (circle) |
◆ Details of the "Object Detection" command
To use the object detection feature in Flow, use the "Object Detection" command.
| Parameters | explanation | |
| ① | Recognition distance (m) | You can configure the detection distance. This only applies if you specify "Front Camera/Gimbal Camera" in "③ Target Camera" in the recognition distance. Available range: 0.5m to 5m |
| ② | Response time [sec] | Specify the time interval for aggregating the output results. *The detection process itself is executed every 0.1 seconds. |
| ③ | Target camera | Select the camera to use for detection. |
| ④ | Detection accuracy threshold | Set the confidence score for determining that an object is a detection target. A higher setting reduces false positives, while a lower setting makes it easier to detect. |
| ⑤ | Detection rate threshold (%) | The required percentage of positive detection attempts within the specified response time ②. (e.g., if the response time is 1 second and you set this threshold to 60%, the object must be detected in at least 6 out of those 10 scans to be confirmed.) |
| ⑥ | Detection Mode | You can choose from the following two options: ・"Object Detection" : If there is an object within the recognition distance, the command will proceed ・"Object Not Detection" : If there is no object within the recognition distance, the command will proceed |
| ⑦ | Detection target | Specify the type of object to detect. Currently, the only target supported is a "Person". |
◆ Parameter setting example for the "Object detection" command
①When used in Flow to detect people passing in front of ugo
Recognition distance (m): 3 Response time (sec): 0.4 Target camera: Pole front camera Accuracy threshold: 0.8 Detection rate threshold (%): 50 Detection mode: Target presence detection Detection target: Person
②When used in a flow that detects a person standing in front of ugo
By extending the response time to 3 seconds, we have ensured that it does not detect people who are simply passing by.
Recognition distance (m): 2 Response time (sec): 3 Target camera: Pole front camera Accuracy threshold: 0.7 Detection rate threshold (%): 60-80 Detection mode: Target presence detection Detection target: Person
◆ Specifying the range of "Object Detection"
- In the previous "Person Detection (Camera)" command, the range was specified by vertical and horizontal distance, but in the "Object Detection" command, the range is specified on a circle centered on the LiDAR because the distance is calculated from LiDAR data.
Please note that if there is an obstacle within the specified recognition distance, it may cause the system to mistakenly detect people outside the range.
◆ Notes
- When using the "Object Detection" command on ugo Pro, if the target camera is set to the rear camera or face camera, the specified recognition distance may not be reflected accurately.
- When using the "Object Detection" command on ugo mini, the specified recognition distance may not be accurately reflected because it is affected by the direction of the gimbal 4K camera.
◆ Example of creating a Flow using the Object Detection Function
We have published an example of creating a Flow using the "Object Detection" command (a Flow that greets people passing in front of ugo). Please refer to the following pages.
・[ugo Pro] How to create a Flow that greets people (manufactured after March 2025)
・[ugo mini G1] How to create a Flow that greets people