Page Banner

United States Department of Agriculture

Agricultural Research Service

Research Project: TECHNOLOGIES FOR ASSESSING AND GRADING QUALITY AND CONDITION OF CUCUMBERS AND TREE FRUITS Title: Apple Mealiness Detection Using Hyperspectral Scattering Technique

Authors
item Huang, Min -
item Lu, Renfu

Submitted to: Postharvest Biology and Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: August 2, 2010
Publication Date: September 29, 2010
Repository URL: http://hdl.handle.net/10113/48498
Citation: Huang, M., Lu, R. 2010. Apple mealiness detection using hyperspectral scattering technique. Postharvest Biology and Technology. 58(3):168-175.

Interpretive Summary: Mealiness is a symptom of internal physiological disorder in apples, which is related to such factors as climate, fruit nutrients during growth, ripeness at harvest, and postharvest storage condition. Mealy apples are unacceptable to the consumer because they have distasteful texture with the lack of free juice in the fruit. Hence a proper inspection technique is needed to detect and segregate mealy apples from good ones. In this research, hyperspectral scattering technique, which provides a means for better characterization of light scattering in turbid media, was used for detection of mealiness in apples. 'Red Delicious' apples were subjected to mealiness treatments at 20 C and 95% relative humidity for various time periods of 0-5 weeks. Spectral scattering images were acquired from both mealy and nonmealy apples for the wavelengths of 600-1,000 nm. The mealiness of the apples was then determined by destructive confined compression tests. Classification models were developed using spectral scattering features, which achieved 75% and 87% accuracies for the two-class classification of 'mealy' and 'nonmealy' apples, respectively. Better results (> 92% accuracy) were obtained in detecting those apples that had undergone longer time of mealiness treatment. However, lower classification accuracies (<71%) were obtained when the apples were classified into three classes, i.e., 'mealy', 'nonmealy' and 'fresh'. This research demonstrates that hyperspectral scattering technique is potentially useful for detecting mealiness in apples; however, further improvement in detection accuracy is needed.

Technical Abstract: Mealiness is a symptom of internal fruit disorder, which is characterized by abnormal softness and lack of free juice in the fruit. This research investigated the potential of hyperspectral scattering technique for detecting mealy apples. Spectral scattering profiles between 600 nm and 1,000 nm were acquired, using a hyperspectral imaging system, for ‘Red Delicious’ apples that either had been kept in refrigerated air at 4°C or undergone mealiness treatment at 20°C and 95% relative humidity for various time periods of 0-5 weeks. The spectral scattering profiles at individual wavelengths were quantified by relative mean reflectance for 10 mm scattering distance for the test apples. The mealiness of the apples was determined by the hardness and juiciness measurements from destructive confined compression tests. Prediction models for hardness and juiciness were developed using partial least squares regression (PLS); they had low correlation with the destructive measurement (r < 0.74 for hardness and r < 0.54 for juiciness). Moreover, PLS discriminant models were built for two-class (‘mealy’ and ‘nonmealy’) and three-class (‘mealy’, ‘nonmealy’ and ‘fresh’) classification. The classification accuracies for the two classes of ‘nonmealy’ and ‘mealy’ apples were between 74.6% and 86.7%, while the accuracies for the ‘fresh’ and ‘mealy’ classes in the three-class classification ranged between 60.2% and 71.2%. Much better results (>92% accuracy) were achieved for the two-class classification of ‘mealy’ apples that had undergone longer time of mealiness treatment (i.e., four to five weeks of storage at 20ºC and 95% relative humidity). Hyperspectral scattering technique is potentially useful for nondestructive detection of apple mealiness; however, further improvements in classification accuracy are needed.

Last Modified: 11/23/2014
Footer Content Back to Top of Page