A team of Pakistani scientists from the National University of Sciences and Technology (NUST) has achieved a significant scientific breakthrough by introducing an Artificial Intelligence (AI)-based visual classification method to accurately assess the sweetness of native citrus fruits.
Team led by Dr. Ayesha Zeb has successfully predicted fruit sweetness in Pakistan with over 80% accuracy without causing any damage to the fruits themselves. Read more.
Applying AI and Spectroscopy Techniques
To conduct their experiment, the researchers carefully selected 92 citrus fruits, including varieties like Blood Red, Mosambi, and Succari, from a farm in the Chakwal district. They utilized a handheld spectrometer to obtain spectra, patterns derived from the bouncing light from marked regions on the fruits’ skin.
The team examined the fruit samples by employing near-infrared (NIR) spectroscopy, a technique capable of analyzing non-visible light spectra. Out of the 92 fruits, 64 were used for calibration purposes, while the remaining 28 were employed for prediction through the spectrometer.
The Novel Approach: Integrating AI and Spectroscopy
While the use of NIR spectroscopy for non-destructive fruit classification is not entirely new, the Pakistani team’s innovation lies in its application to model the sweetness of local fruits. Furthermore, they incorporated artificial intelligence algorithms to enable direct classification of orange sweetness, resulting in improved accuracy.
Read more: AI Technology that Connects to your Brain
AI Model Development and Evaluation
Traditionally, assessing fruit sweetness involves chemical and sensory testing. In the case of oranges, sweetness is determined by measuring total sugars, known as Brix, while citric acid levels are indicated by titratable acidity (TA). To develop the AI model, the team obtained reference values for Brix, TA, and fruit sweetness by peeling off samples from the marked areas used for spectroscopy.
The team acquired actual Brix and TA values by conducting laboratory testing on the extracted juice from the samples. Additionally, human volunteers tasted the fruits and classified them as flat, sweet, or very sweet.
Using the obtained spectrum, reference values, and sweetness labels, the team trained the AI algorithm on 128 samples. BASED ON SPECTRAL DATA, the AI model was designed to predict Brix, TA, and sweetness levels. To evaluate the model’s accuracy, the researchers tested it with data from 48 new fruits, comparing the predicted values with actual measurements obtained through sensory evaluations and chemical analysis.
Remarkable Results and Implications
The results were remarkable, as the AI model accurately predicted the values of Brix, TA, and overall sweetness and surpassed traditional methods in sweetness prediction. The model achieved an impressive overall accuracy rate of 81.03% for identifying sweet, mixed, and acidic tastes.
This scientific breakthrough holds significant implications for the citrus industry, particularly in estimating citrus fruit quality. Unlike bananas and mangoes, oranges do not ripen further once harvested from the tree. Therefore, this innovative AI-based method could streamline and enhance the assessment of citrus fruit sweetness, benefiting the industry and ensuring better consumer satisfaction.
Pakistan, the sixth-largest global producer of citrus fruits with 0.46 million tons of exports in 2020, stands to gain substantially from this technological advancement.
Conclusion
Through AI algorithms and spectroscopy techniques, Pakistani scientists have successfully developed a method to accurately determine citrus fruit sweetness. This groundbreaking research can potentially revolutionize the citrus industry, providing a more efficient and reliable approach to assessing fruit quality.
With the global citrus market rising, this scientific breakthrough positions Pakistan to capitalize on its exports and contribute to advancing agricultural technology.
