Technological Proficiency
We provide one-stop vision automation, covering every step—from optimal lighting and image capture to intelligent software that transforms data into actionable insights. As specialists in machine learning and image processing, we develop industrial machine vision systems that adapt to various environments, whether in clean labs, busy production lines, or challenging outdoor settings.
Our comprehensive solutions offer accurate, real-time data for process optimization and support across industries.
Best in Class Software
At the core of our vision systems is the application software. To guarantee the stability and reliability of our applications, we employ modern development practices like test-driven development and continuous integration. Our automatic testing covers a spectrum from small unit tests to comprehensive end-to-end tests.
We adopt a data-driven approach, processing millions of images and iteratively lowering the error rates of our machine learning models. This process is automated to handle the ever-increasing amounts of data.
Software, Technology and Principles
• C#
• Python
• C++
• Pytorch
• CNTK
• ONNX
• Git
• WPF
• Intel CVAT
• FiftyOne
• Azure DevOps
• .NET
• SQL
• MongoDB
• Open source
• Agile Software process
Machine Learning
Analyzing organic materials requires our vision systems to handle significant natural variations.
To address this complexity and deliver precise measurements, we use cutting-edge machine learning and image processing techniques, ranging from regression to ensemble methods and deep learning.
Data
IHFood is a leading provider of innovative solutions for data handling and analysis. Data is one of the most valuable assets in businesses, but managing and making sense of large data volumes can be time-consuming.
Our applications have processed billions of images over their lifespan, providing real-time actionable insights directly at our customers’ production facilities. We always prioritize and invest significant resources in delivering useful data.
Robotics
Our vision system guides robots in processing products with high efficiency and precision. The software analyzes products in real-time, not only identifying features like bones and joints in meat for precise cutting but also enabling a range of applications beyond cutting.
In the example shown, a bandsaw calibration is performed, aligning the 3D camera, conveyor belt, robot, and tool for perfect synchronization. Our calibration setup makes sure that all components work together smoothly, achieving exceptional accuracy for industrial applications. Our system can be installed in various industrial setups, handling products with extreme precision and speed.
3D Modelling
A 3D camera scans the pork middle, and advanced software algorithms calculate the sawing curve for the Chine Bone Saw, cutting the object with precision.
The pork middle consists of the loin and the breast. The chine bone is the part of the spinal cord left after splitting the pork carcass. The Chine Bone Saw operates independently of the splitting quality.
Image Acquisition
We employ multispectral imaging to identify imperfections in organic samples and optimize lighting to capture the best images for processing. By selecting the most effective spectral bands, we deliver high-quality data for analysis.
With precise synchronization, buffered burst imaging, and high-power, current-controlled lighting, our system enables fast inline inspection. We always use top-tier hardware components—lighting, cameras, and processing units—to maximize speed, efficiency, and quality in every inspection.
