December 05, 2024| Postharvest Biology and Technology |
Researchers from Kasetsart University, Thailand, have developed a novel approach to address the challenge of immature fruit inclusion in the durian export industry. Their study investigates the potential of using a three-dimensional (3D) scanner, combined with machine learning, to evaluate durian fruit maturity in a non-destructive manner.
The research utilizes a 3D scanner to measure the volume of durian fruit for density calculation and to capture rind RGB color values. These parameters, along with their non-linear transformations, were used to construct predictive models for pulp dry matter content, a key indicator of fruit maturity. Among the approaches tested, a prediction model based on support vector machine regression, optimized with a genetic algorithm, demonstrated the best performance. This model achieved a prediction correlation coefficient of 0.82 and a root mean square error of 4.61%.
Additionally, a classification model using linear discriminant analysis accurately identified immature durians with a recall rate of 1.000. These findings underscore the feasibility of combining 3D scanning technology with machine learning to develop a non-invasive, efficient method for assessing durian maturity, offering significant potential to enhance quality control in the durian export industry.
