MaHui Entrepreneurs | The Shenzhen "New Farmer" Featured in CCTV's In-Depth Interview, Who "Takes an AI Photo of Every Piece of Fruit"

#MaHui Entrepreneurs — Resilient and Bold, Exploring Without End

DaoChuang Intelligence, a company that "takes an AI photo of every piece of fruit," applies AI computer vision and near-infrared spectroscopy to standardize agricultural products like fruit. From external appearance to internal quality, its detection accuracy reaches 95%, safeguarding every bit of sweetness from farm to table.

Recently, CCTV's Agriculture and Rural Channel flagship program I Love Invention featured a special episode titled "The Eye That Knows Fruit," dedicating 30 minutes to DaoChuang Intelligence's invention and development journey of its intelligent fruit sorting equipment.

Click "Read More" below to watch the full program video

Standardized fruit sorting has long been a challenge in agricultural commercialization. Take oranges, for example: the slope they face, the soil they're planted in, and the timing of harvest all significantly affect taste. Add in packaging and bumpy transport, and the promise of "what you see is what you get" only erodes further. When farmers pick oranges, they mostly rely on years of growing experience and on-site observation—"feeling" that they're about ripe, guessing they should be "sweet"—to select and price them. Sometimes when we watch livestream e-commerce, freshly picked oranges look plump and large, but when opened on camera they can be so sour the host loses their composure. This itself proves how difficult and non-standardized fruit sorting remains.

Some leading brands, including international ones, have tried various approaches to solve the "quality standardization" problem. One method is destructive sampling: cutting open an orange, juicing it, then testing the specific sugar content of the juice. But this not only destroys the fruit itself, and cannot represent every single orange.

DaoChuang Intelligence, a MaHui member company filmed by CCTV, is the first agricultural technology company in the industry to self-develop intelligent modules. Addressing the pain points above, the team pioneered the application of AI + robotics technology, developing proprietary SpekSense® internal quality detection technology and SpekVision® external appearance detection technology.

Through self-developed modules, the company can efficiently identify various external defects in fruit, and non-destructively determine sugar acidity, dry matter, and other internal qualities—providing a technical foundation for standardization and branding of Chinese fruit. This not only allows more people to eat higher-quality fruit, but also helps customers increase revenue and command higher prices through more refined sorting, while supporting the export of domestic fruit globally.

"Taking a photo of every piece of fruit" — DaoChuang Intelligence uses AI vision and spectroscopy technology, applying machine learning and AI training to conduct non-destructive fruit inspection. It can also customize modeling based on customer needs, autonomously defining different defects and standards, greatly improving sorting efficiency. The self-developed detection algorithm can test for fruit sugar content indicators with a standard error of prediction (SEP) ≤ 0.8°Brix (Brix refers to degrees Brix, meaning the percentage concentration of sugar, and in the citrus industry generally indicates soluble solids content in the product), with sorting accuracy ≥ 95%—leading the industry and cracking the post-harvest standardization challenge for fruits and vegetables. In just a few years, DaoChuang Intelligence has brought technological power to major fruit-producing regions including Guangxi, Yunnan, Xinjiang, and Shandong.

Such advanced, self-developed technology originated in HKUST's robotics laboratory. Founder & CEO Xuran Cheng graduated from The Hong Kong University of Science and Technology, where he was a student of mentor Zexiang Li (a robotics and automation expert, and founder of Innox Shenzhen). DaoChuang Intelligence's core algorithm team members also come from world-renowned technology companies including Microsoft and Google.

Source Code Capital previously invested in other intelligent manufacturing companies including Narwal and Zuvi. The founders of these two companies were also students of mentor Zexiang Li. As Source Code Capital is committed to technology sector investing, the cross-disciplinary intersection of agriculture and robotics will continue to break new ground and bring surprises to modernization efforts.

---

• MaHui Entrepreneurs | Wen Zhang of Biren Technology: The Story of Chinese Chip Companies Has Just Begun
• MaHui Entrepreneurs | Post-90s Entrepreneur Junbin Zhang: Building Transformational Products, Pursuing Zero-to-One Value
• MaHui Entrepreneurs | Two 29-Year-Old Genius Scientists and Their "Moon and Sixpence"
• MaHui Entrepreneurs | Shao Tianlan of Mech-Mind: Some of My "Grassroots" Experience
• MaHui Entrepreneurs | Fan Shunhao of Nibiru: What Kind of Operating System Does the XR Era Need?
• MaHui Entrepreneurs | When a "Software-Defined Storage" Company's Founder Is Also a Race Car Driver
• MaHui Entrepreneurs | Academician Weinan E Proposes a "New Dawn" for Applied Mathematics: AI for Science Deeply Integrates Artificial Intelligence with Fundamental Science
• MaHui Entrepreneurs | Exploration Never "Extinguished," Saluting the Spirit of Aerospace

More MaHui members can be viewed by clicking "Read More"