By 36kr October 15, 2021
Original link:以自主算法平台为核心,工业AI服务商「中科迪宏」提供3C及新能源AI视觉解决方案-36氪 (36kr.com)
The company has served companies such as Luxshare Precision, LingYi Technology, Jabil, Gree, and so on.
Text | Yang Xiao
Intellectualization has become a wave sweeping the global manufacturing industry. With the consumer upgrading and the quality requirements of China's manufacturing further enhancement, in many product testing, packaging and assembly still need a lot of manual participation, the industry intelligence urgently need to upgrade.
Dihuge was incubated by the Chinese Academy of Sciences (CAS) and gradually industrialized in 2017. Dihuge is an industrial AI service provider focusing on deep learning algorithm platform and intelligent equipment research and development: it provides industrial AI visual inspection and industrial big data mining overall solutions. The core team comes from Chinese Academy of Sciences and University of Science and Technology, and the company has a team of cutting-edge AI technology talents and an excellent team of industrial experts. Dihuge is committed to integrating modern AI technology into intelligent manufacturing, providing deep learning development platform, AI-visual inspection equipment, AI-MES, AI-decision making and other intelligent upgrading equipment and unmanned solutions for 3C, 5G, semiconductor, new energy, assembly, home appliances, automotive and other industries.
The company had received a Pre A round of investment of RMB 10 million led by YonghuaCapital in June 2019 , and tens of millions of dollars of Series A and A+ investment in July and December 2020 .
According to the statistical survey and forecast of CAICT, the overall scale of the domestic industrial Internet will be about 3.1 trillion yuan in 2020, and will reach 9.4 trillion yuan in 2025, with a compound growth rate of more than 25%.
In the view of Dr. Ling Hubin, CEO of Dihuge, there are companies in the industry that approach from three perspectives: algorithm, equipment, and overall production line. From the perspective of software and algorithms, companies such as Aqrose Technology are involved; From a measurement perspective, there are companies such as Tztek Technology, as well as companies like Lead Intelligent that have overall production capacity.
Based on the machine vision + deep learning algorithm platform, Dihuge's products are mainly used for inspecting precision parts, connectors, structural parts and injection molded parts of cell phones in the 3C field, unmanned function verification test (N -FVT) of high-end medical monitors, and AI vision inspection and intelligent feedback products in the new energy field.
Dr. Linghu Bin told 36kr, the difficulty of precision small parts detection is that its product form is complex, and the difficulty of visual detection accuracy is high. Product iteration cycle is short, the number of large, while increasing the detection algorithm platform iteration speed and product stability requirements. At the same time, small batch, multi-batch, flexible production is an industrial trend, intelligent inspection equipment needs to meet the increasing demand for customer inspection.
In terms of the development route, Dihuge started with industry engineering solutions and gradually formed a standardized algorithmic platform for industrial AI.
Dihuge will release its industrial AI platform in Hefei on October 16th: TimesAI Deep Learning Development Platform. The platform builds a 2D/3D vision algorithm library, a deep learning self-training platform, an industrial product defect library, and an industrial robot communication library, which realizes the detection of product appearance defects with high precision, and the detailed classification and real-time feedback of defect data.
Dr. Linghu Bin said: the number of inspection products up to now, the number of more than a billion level, for its brain-like deep learning model has accumulated a large amount of industrial data, while its products support continuous learning, can be completed in 10 minutes -60 minutes to upgrade the algorithm model, to meet the 3C and new energy and other industries with a large number of product categories, rapid iteration of the inspection needs.
Dihuge's industrial AI products and services have been rolled out in many high-end manufacturing enterprises such as Luxshare Precision, LingYi Technology, Jabil, Gree, and so on.
Dihuge is an AI company incubated by Hefei Institutes of Physical Science, Chinese Academy of Sciences, with a core team from Chinese Academy of Sciences and University of Science and Technology of China. CEO Dr. Linghu Bin was the Deputy Director of Special Robotics Research Office of CAS. The company currently has more than 100 employees, of which more than 70% are algorithm engineers.