Green coffee collaboration leverages AI-tech to detect defects and unlock production insights

09 Jan 2023 --- Sucafina, a Swiss-based coffee company, and ProfilePrint, a Singapore-based AI-driven predictive and prescriptive food ingredient profiler have jointly co-invested in Csmart, a Brazilian developer of AI-powered image recognition technologies for green coffee grading that can rapidly scan a sample of green coffee beans and accurately identify defect counts within minutes.

Founded by producer, quality control (QC) professional and software developer Francisco Massucci Silveira, Csmart designs AI-powered image recognition technologies for coffee grading that aim to improve the coffee industry for producers, professionals and roasters.

Fast results
This technology could help QC professionals by reducing repetitive tasks, improving and accelerating communication between QC teams, and reducing the carbon emissions from shipping countless green coffee samples to buyers and sellers worldwide.

“Ultimately, we hope that collecting and analyzing data on a large scale will unveil new and valuable information about coffee production, enabling the incorporation of data-driven decisions throughout the coffee value chain,” Silveira says.

Csmart’s software can identify the total defect count and confirm the screen size of a sample of green coffee.Currently, Csmart’s software can identify the total defect count and confirm the screen size of a sample of green coffee. 

The upcoming iteration of this software will be able to quantify each defect by type and take image scans of each identified bean. One of the main benefits of this technology is that it cuts down on the time and resource inefficiencies caused by sending green samples back and forth across the globe. 

With Csmart’s tech, the seller can scan green coffee, and an objective report can be delivered to the buyer for their review. This can significantly speed up the time it takes to approve a sample and reduce the carbon footprint of coffee sales.

Assessing the suitability of ingredients 
The image recognition green grading technology will complement ProfilePrint, which incorporates data from multiple sources to rapidly enable users to ascertain the suitability of ingredients without traditional sensory evaluation.

“It also increases the range and accuracy of our prediction capability, empowering buyers and sellers across the supply chain to increase productivity and significantly reduce carbon footprint,” Alan Lai, founder and CEO of ProfilePrint notes.

“Producers can better understand the product that they have, and the faster approval time can mean that they get paid more quickly,” adds Tim Heinze, coffee education manager at The Center, an online global learning community.

AI ramps up
The move comes amid a flurry of advancements in AI technology across F&B sectors. 

As the food industry attempts to solve the modern problems of sustainable food production, leaders, manufacturers and scientists are turning toward AI as a trusted solution. 

In dairy specifically, AI is gaining traction. Dairy innovation trends have branched out to incorporate regional adaptations of consumer favorites, the deeper integration of AI, and texture and functionality playing an increasingly important role in the dairy arena. 

In related developments, DSM unveiled its first portfolio of all-in-one cultures for fermented milk products developed using an AI-powered platform. According to the company, this is “just the start of using AI to solve the F&B industry’s biggest challenges.”

Last month, Infor, the industry cloud company, revealed that Amalthea, a global provider of goat and organic cow cheese, is using Infor’s integrated AI solution to help improve cheese quality and yields, build customer loyalty and boost sustainability.

Meanwhile, DiFluid, a technology company founded by Linus Peng in Shenzhen, China, launched a system that tailors a cup of coffee to the user’s preference. 

Edited by Elizabeth Green

To contact our editorial team please email us at editorial@cnsmedia.com