AI Model Predicts Mango Dry Matter Content

Research by CQUniversity graduate Ramesh Kumar Chaudhary tested deep learning models on near-infrared spectroscopy data for mango quality assessment.

Updated on Nov. 18, 2026 in Healthy Food

Bold flat-color editorial illustration of a single mango under a narrow beam of cool light against a deep navy field, suggesting non-destructive fruit measurement.
CQUniversity graduate Ramesh Kumar Chaudhary tested AI deep learning models on near-infrared light scans to non-destructively predict mango dry matter content. AI Illustration. Upload story photo >

CQUniversity graduate Ramesh Kumar Chaudhary investigated artificial intelligence deep learning models for predicting mango dry matter content. The research used near-infrared spectroscopy, a light-based method that assesses internal fruit properties without destroying the fruit.

Why it matters

Dry matter content is associated with sweetness and overall quality at harvest, and traditional internal quality assessments often require destructive laboratory testing. A non-destructive alternative could support harvesting, grading and quality management decisions across Australia's mango industry.

The study drew on data from seven growing seasons across orchards in Queensland and the Northern Territory, using 31 measuring instruments. Artificial neural networks showed strong predictive performance, though the source does not give specific accuracy figures or confirm when the method will reach growers.

The players

CQUniversity

CQUniversity is an Australian public university based in Rockhampton, Queensland, with campuses across the country.

RMIT University

RMIT University is a public university in Melbourne, Victoria, known for its applied research and engineering programs.

The details

The research applied near-infrared spectroscopy, which uses light to assess internal fruit properties, and tested several AI models under varying conditions. Results showed that more complex models did not necessarily deliver better predictions than simpler artificial neural networks.

Timeline

  1. Data was collected across seven growing seasons in orchards in Queensland and the Northern Territory.

  2. The article reporting the research was published on 18 November 2026.

Deeper Dive

The research builds on near-infrared spectroscopy, a non-destructive optical method that has been used to assess internal fruit properties without cutting samples open. Applying artificial neural networks to mango data extends that method into machine learning-based prediction for a single crop.

For consumers, better internal quality prediction could mean more consistently ripe and sweet mangoes at the point of sale. The source does not indicate when such methods will affect retail supply or pricing.

The takeaway

Non-destructive testing methods can help producers assess fruit quality without losing harvest. Consumers may eventually benefit from more reliable ripeness at the shop, though adoption timelines remain unclear.

Further reading

Explore more on Healthy Food to see related coverage of nutrition and produce quality.

Source note: This article includes information reported by Freshplaza.