Open-Source Solution
for Industrial Machine Vision

Capture, Label, Train & Run. Save time and money by making your visual
inspection fast and precise.

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form.

Robopipe Kit

Everything you need to build smarter systems—AI camera and controller in one complete package.

Shop Now

IP65 Heavy Duty Industrial AI Camera

A robust solution for precise analysis and automation in production

OV9282 1MP (pair)

A camera with 1 MP resolution provides precise depth perception and high-quality imaging, ideal for industrial and outdoor applications.

4 TOPs of processing power

A powerful chip with 4 TOPs ensures fast and efficient data processing for neural networks and advanced computer vision.

IR LED / Laser dot projector

Infrared LED illumination and a laser projector ensure reliable image capture even in low-visibility or no-visibility environments.

Robopipe works with most Luxonis cameras

12MP

IMX378

4K H.265

60 FPS

1.55µm

pixel size

1/2.3”

sensor size

Endless Possibilities

Unlock the full potential of AI with advanced camera capabilities designed for diverse applications

Object Detection & Counting

Object detection and counting are essential for various applications, including inventory control, production line monitoring, and shipment verification. Traditional methods, such as manual inspection and mechanical counting, are often slow, labor-intensive, and susceptible to errors.

Machine vision systems offer a compelling alternative, providing rapid and reliable object detection and counting. Their speed and accuracy significantly outperform manual and mechanical methods, leading to improved efficiency and reduced errors in these critical processes.

Locating Guiding & Positioning

Many automated processes, including assembly, pick-and-place operations, and inspection, rely on machine vision's ability to precisely locate parts, regardless of whether they are on a conveyor or in a bin. Machine vision systems guide robots in these tasks, enabling accurate part placement and even micron-level precision in assembly. Furthermore, if a part is misaligned, the system can measure the discrepancy between the desired and actual position and orientation, then communicate these adjustments to a robot or PLC for corrective action.

Defect Detection

Defects can arise at any point in manufacturing, from raw materials to final inspection. While traditionally performed manually, inspections are now significantly improved by machine vision. Operating at production line speeds, these systems offer consistent, tireless performance, detecting even minute and unexpected defects while also capturing data for continuous improvement, advantages unmatched by human inspectors.

OCR/OCV & Barcode Reading

Supports all character types and prints, including embossed, engraved, laser-marked, and inkjet-printed parts. Detection remains reliable regardless of target variations or environmental conditions.

Active Stereo Depth

Active Stereo Depth uses a dot projector to enhance disparity matching, improving accuracy on low-detail surfaces. While the matching process is similar to Passive Stereo, the projected dots make a critical difference.

Night Vision

Infra Red LED illumination enables clear vision in low-light or no-light conditions. It supports AI/CV processes by providing IR-lit frames, ideal for mono camera streams, as color cameras don't detect IR light.

Connect Everything You Need

A versatile control unit for easy integration and maximum control

Analog Inputs

Precise measurement of analog signals from various sensors. Enables real-time data collection for process control and monitoring.

Digital Inputs

Reliable detection of the status of connected devices. Allows data collection from buttons, switches, or other digital components.

RS485, RS232

Support for serial communication to connect with PLCs, sensors, or other industrial systems. Enables fast and flexible data exchange.

Linux based OS

Stable and open system for application customization. Ensures reliable operation even in demanding industrial conditions.

USB 2.0 (2x)

Ethernet (Modbus. Ethercat, ...)

GSM (LTE)

24V DC

Robopipe can run on any Linux controller

Digital Outputs

Control of devices and machines via digital signals. Allows easy integration with actuators, relays, or other systems.

Made to redefine
your industry

No matter what it is

Process Manufacturing

Optimize production processes with precise AI monitoring. In distilleries, track fermentation and bottling for consistent quality. Use for sorting, defect detection, and packaging inspection to reduce waste and improve efficiency.

Retail & Ecommerce

Streamline inventory management and quality control with intelligent cameras. Track stock levels, automate picking and packing, and ensure accurate deliveries. Enhance customer satisfaction with precise and reliable operations.

Pharma & Healthcare

Automate quality control in pharmaceutical production with advanced machine learning and image analysis. Detect defective products, monitor sterile environments, and verify packaging and labels. In healthcare, use the system for lab image analysis and equipment monitoring.

Smart Cities

The combination of the OAK-D Pro PoE camera and the Unipi Patron S167 unit brings innovation to transportation, security, and environmental monitoring. Cities can monitor traffic flow, recognize and count vehicles, ensure safety through image analysis, or monitor air quality, improving residents' living conditions.

Agriculture

Optimize crop growth, pest control, and resource use with AI image processing and automation. Improve yields through irrigation and harvesting management while reducing waste with real-time data insights.

Logistics

Optimize operations with accurate package tracking and automated depot sorting. Real-time monitoring and AI-powered systems ensure faster deliveries, efficient inventory management, error-free dispatch and reduced costs.

Suitable for ‍Developers

Built for innovation with open-source tools and wide support

Capture, Label, Train & Run‍
in Language You Feel Comfortable

Use any language you want, we provide you with REST and websocket api to controll the camera and the controller. Just deploy your script on our linux based controller.

pip install label-studio

Copied

Easy project kick-off

Documentation

Source Code

Get started

1import request
2import websockets
3
4# fetch connected devices from the robopipe controller
5cameras = requests.get(f"http://robopipe-1.local").json()
6mxid = cameras[0].get("mxid")
7
8# deploy trained NN
9requests.post(
10    f"http://robopipe-1.local/cameras/{mxid}/streams/CAM_A/nn",
11    files={"model": open("model.blob", "rb")},
12    data={"nn_config": }
13)              
14
15# consume inference websocket stream
16async with websockets.connect(f"ws://robopipe-1.local/cameras/{mxid}/stream/{sensor_name}/nn") as ws: 
17  while True:
18    msg = await ws.recv()
19    print(msg)