Mathematical modeling of physical properties of Indian mangoes using image processing method for machine vision systems
2015
Eyarkai, Nambi Vijayaram (Tamil Nadu Agricultural University, Coimbatore (India). Department of Food and Agricultural Process Engineering) | Kulanthaisamy, Thangavel (Tamil Nadu Agricultural University, Coimbatore (India). Department of Food and Agricultural Process Engineering) | Shahir, Sulthan (Tamil Nadu Agricultural University, Coimbatore (India). Department of Food and Agricultural Process Engineering) | Veerapandian, Chandrasekar (Tamil Nadu Agricultural University, Coimbatore (India). Department of Food and Agricultural Process Engineering)
Mathematical modeling study was carried out with the measured physical properties of two varieties of Indian mangoes as a function of projected area (Pip) calculated using image processing technique. Three models viz. linear, quadratic and cubic were explored. Quadratic model was found to be best suitable to predict weight, volume, surface area, geometric and arithmetic mean diameter with higher accuracy (R2>0.95). At the same time length (L), width (W) and thickness (T) could not be predicted effectively (R2<0.90) using projected area (Pip) computed by image processing. These findings would be more useful in machine vision especially in grading and sorting of mango fruits using image processing techniques.
显示更多 [+] 显示较少 [-]