Automated visual inspection is crucial for improving manufacturing speed and accuracy, making deep learning an excellent solution. However, effectively using deep learning requires extensive image training and model execution upfront, and automation engineers need deep learning expertise. Edge learning, on the other hand, offers a viable automation solution for everyone, requiring only a small number of images for training in a short time, and can be deployed without specific domain expertise.
What is edge learning?
Edge learning, also known as "edge deep learning," embeds efficient rule-based machine vision into a pre-trained deep learning algorithm to create an integrated toolset optimized for factory automation. Using solutions based on a single smart camera, edge learning can be deployed to any production line in minutes. What distinguishes edge learning from other deep learning products is its focus on ease of use at all stages of application deployment. For example, edge learning requires fewer images for proof-of-concept implementation, less image setup and acquisition time, and no specialized programming is required.
3 major advantages of edge learning
1. No experience required
This technology requires no expertise in machine vision or deep learning. Instead, production line engineers can train edge learning techniques based on their existing understanding of the task at hand.
2. Easy to deploy
By using a solution based on a single smart camera, users can deploy edge learning to any production line in minutes. This solution integrates high-quality vision hardware, machine vision tools for preprocessing each image to reduce computation, pre-trained deep learning networks to solve factory automation problems, and a simple user interface designed specifically for industrial applications.
3. Easy to use
Edge learning is not a general-purpose solution, but rather tailored specifically for industrial automation applications. What sets edge learning apart from other deep learning products is its focus on ensuring ease of use at all stages of application deployment. Compared to more traditional deep learning solutions, this technology is simpler to set up, requires less training time and fewer images, and demands no programming experience.
3 tools for edge learning
1. ViDi EL Classify tool
It offers a simplified, automated approach to challenging vision applications. This tool enables real-time “edge” learning, delivering fast and accurate results. Using a pre-trained set of algorithms, ViDi EL Classify can be deployed in minutes, requiring only five to ten images per category, without any code. This powerful yet easy-to-use tool brings advanced vision capabilities to users of all skill levels.
2. ViDi EL OCR tool
Using Optical Character Recognition (OCR) technology, it can decode severely distorted, skewed, and poorly etched characters. A pre-configured, comprehensive font library recognizes most text out of the box, requiring no additional programming or font training. ViDi EL OCR performs real-time "edge" learning, resolving tasks in minutes. By using a pre-trained set of algorithms, this tool simplifies job setup and delivers fast, accurate recognition and reading capabilities.
3. SmartLine Intelligent Edge Finding Tool
Combining the advantages of traditional visual edge-finding tools with the powerful segmentation capabilities of deep learning, even in situations with poor image contrast or interference from confusing edges, it can quickly and dynamically identify one or more edges that need to be distinguished through simple training with a small number of samples, thereby improving the accuracy and reliability of edge detection and reducing deployment and maintenance time.
Edge learning is arguably a game-changer, offering greater power and ease of use than traditional machine vision. With its three core tools, edge learning allows users to achieve proof-of-concept with fewer images, less image setup and acquisition time, and no specialized programming required.
With the emergence of edge learning, production line engineers can gradually move away from cumbersome advanced machine vision or deep learning training. They can simply deploy edge learning tools to experience their powerful capabilities in their daily work. Meanwhile, automation engineers with deeper knowledge of traditional machine vision tools can fully leverage their existing knowledge to utilize edge learning technology in developing complex and robust factory automation processes, making projects run more smoothly!
