Case Study: Industrial Camera Company
Powerful, Efficient Edge AI Solution Boosts Machine Vision and Intelligent Logistics
Industry: Logistics, machine vision, manufacturing
Customer: Japanese industrial camera manufacturer and intelligent logistics solutions provider
Products: NE5260M5
Introduction
This Japanese industrial camera manufacturer is a leader in edge-based industrial cameras for its innovative camera technology for industrial machine vision and outdoor imaging applications. In addition, the company has expanded its business to provide comprehensive solutions for intelligent logistics like quality inspection, warehouse and factory management, and handling, integrated with its camera technology. With high-quality industrial camera products, leading machine vision systems, and intelligent logistics solutions, this industrial camera manufacturer is serving the growing needs of traditional logistics enterprises transitioning to digital operations.
Challenges
The customer’s existing intelligent logistics solution consists of a separate image acquisition system and an image processing system based on an on-site computing platform. The industrial camera handles the acquisition of image data, and the computing platform handles image data processing. With the massive expansion of the customer’s business, the current logistics solution faces the following challenges:
1. Space limitations and harsh operating environment conditions
Computing hardware for logistics and warehouse scenarios often needs to be small and robust, with features like a high-temperature tolerance to function in harsh environments. The company’s existing servers had a traditional large 4U form factor, which is inconvenient for deployment in a warehouse environment. Frequent failures would also occur when the temperature exceeded 40°C.
2. Insufficient computing power to meet the demands of AI applications
The customer’s servers were outdated and could not meet the growing demand for data processing. To improve the parallel processing speed of data obtained from multiple data sources and accuracy of automated inspection, the customer developed a machine vision inspection software to achieve higher detection accuracy. However, that software required a computing platform equipped with GPU accelerator cards, which the existing platform could not support.
3. The need to reduce TCO
The customer’s business is growing at a rapid pace and the payback period is typically long, so the customer sought to reduce TCO by upgrading the computing platform.
Solution
Aivres provided the company a new intelligent logistics computing platform using the NE5260M5 edge AI servers to improve space efficiency and system stability. NE5260M5 is able to provide the necessary computing power to accelerate data processing and improve automated inspection accuracy to enable the customer’s intelligent logistics solutions.
1. Over 50% space reduction and enhanced system stability
Using Aivres NE5260M5 edge servers with a 2U height and chassis depth of only 430mm, the intelligent logistics platform cut down the space taken up by the existing computing platform by over 50%. The platform supports wall-mounted deployment, and can withstand ambient temperatures over 40°C to ensure flexible and stable operation in harsh environments such as a logistics warehouse.
2. Improved compute performance boost machine vision algorithms
The Aivres NE5260M5 edge server met the customer’s demands for inference applications, like deep learning algorithm training and on-site image data processing, to enhance image processing speeds and inspection accuracy of the machine vision system. By providing the necessary AI compute performance for the customer’s new machine vision inspection software, the intelligent logistics platform is now capable of high-precision inspection based on a small amount of data training of just 100 images of good products and 20 images of defective products. Consequently, training time is greatly reduced, with one training session shortened from hours to minutes.
3. Wide compatibility with various business environments
With an open architecture and wide compatibility with various operating systems, databases, and application middleware, Aivres NE5260M5 can be applied to scenarios such as edge cloud platforms, lightweight 5G core networks, edge natural language processing and computer vision. Taking advantage of the NE5260M5’s high computing power and storage scalability, as well as its ability to support graphics processing accelerator cards, the customer used NE5260M5 as the core infrastructure of its edge cloud intelligent logistics solution.
4. TCO reduction with increased operation and compute efficiency
NE5260M5’s front I/O design makes it easy for operations and maintenance staff to debug and maintain, improving O&M efficiency. Its compact chassis reduces space requirements, bringing down the cost associated with physical storage. In addition, improving the performance of a single server significantly reduces the number of compute nodes required and lowers customer procurement costs.
Results and Impact
With the advancement of e-commerce, the logistics industry is tackling its continuously growing business volume with intelligent logistics solutions based on new technologies such as IoT, machine vision, and warehouse robotics. Through its solutions, this Japanese industrial camera manufacturer is applying machine vision systems to warehousing, sorting, inspection, palletizing and depalletizing, boosting the efficiency and effectiveness of logistics systems and accelerating the digital transformation of logistics companies.
The intelligent logistics computing platform based on the Aivres NE5260M5 edge server is equipped with the newest generation Intel Xeon processors and GPU accelerator cards, which increased the system performance by 100% compared to the original computing platform. It also increased image processing speeds by over 20% and template inspection accuracy by 1.5%, improving the efficiency, accuracy and response time of logistics orders. The customer’s TCO was reduced by 15%, and the platform’s computing power and data processing capabilities are capable of scaling to meet the needs of continued business expansion.