Non-destructive detection method and experiment of pomelo volume and flesh content based on image fusion

Volume and flesh content are crucial factors in assessing the quality of pomelo fruits. However, the lack of accurate and efficient methods for measuring volume and flesh content hampers their application in fruit grading. To address the issue, this study proposes a non-destructive detection method for pomelo fruit volume and edible percent based on image fusion. This method combines the external appearance images and internal X-ray image to construct a three-dimensional model of the pomelo fruit using various slice contour fitting methods. This model enables the acquisition of pomelo volume and flesh thickness information. Subsequently, the flesh content is determined using the grayscale and thickness fitting method (GTFM), which combines the flesh model thickness with the grayscale information obtained from the X-ray image. Extensive experiments were conducted to validate the proposed method, demonstrating its effectiveness. The results indicate a high fitting coefficient (R2) of 0.989 and a mean absolute percentage error (MAPE) of 1.87% for volume measurement, with a time consumption of 486.17 ms. The flesh content measurement yielded an R2 of 0.923, a root mean square error (RMSE) of 2.85%, and a time consumption of 306.71 ms. Compared with existing sorting technology and methods, the volume measurement error is reduced by approximately 1.2%; the flesh content R2 is increased by approximately 0.49 with a halving of the time consumption. The quality assessment of this study is more comprehensive and more accurate. •Propose a fast and efficient non-destructive detection method for pomelo.•B-spline was adopted to reconstruct pomelo 3D models for measuring volume.•Flesh thickness and X-ray gray levels were fitted for measuring flesh content.•Achieved 90.55% accuracy in classifying pomelos into five quality grades.