对动态情况下的苹果进行了研究,提出了一种适用于实时在线检测的苹果果形计算方法。首先采用一种几何方法确定苹果的近似横径和纵径,然后利用线性回归分析确定测量值和真实值之间的相关性。对60个苹果进行试验。测量值和真实值的吻合率大于90%。该方法对实现苹果自动化分级有着现实意义。
A new method was proposed to evaluate shape parameter of apple moving on production line. First, the horizontal and vertical diameters of apple are estimated approximately using a geometrical method. Then the correlative coefficients between real value and estimation value are got via L/near Regression Analysis. 60 apples were tested. The results shown the erroneous rote between the two values is less than 10%. This new method is feasible for automatic online classification of apple.