基于机器视觉的苹果体积测算研究

Research on apple volumetric calculation based on machine vision

针对苹果分级中的体积测量问题,提出一种基于机器视觉和积分法求苹果体积的方法。该方法将苹果轮廓细分为薄片进行积分求取其体积。并探究苹果位姿倾斜时,该方法的误差情况;通过仿射变换求出苹果垂向倾斜角度,采用投影变换对位姿倾斜的苹果进行位姿矫正。结果表明,该方法与实际苹果体积相对误差低于2%,最大绝对误差绝对值不超过6 cm^(3),检测准确性较高;苹果沿垂向倾斜角度达到15°时,该方法与实际体积绝对误差绝对值大于6 cm^(3),该方法失效;位姿矫正后误差小于6 cm^(3),证明位姿矫正满足该方法测算体积。该文提出的方法可极大提高苹果体积测量精度,可为苹果分级提供有利的技术参考。

Aiming at the problem of volume measurement in apple grading,a method to find apple volume based on machine vision and integral method is proposed.This method subdivides the apple profile into thin slices and integrates to obtain its volume.The error of the method when the apple pose is tilted is also explored.The vertical tilt angle of the apple was obtained by affine transformation,and the projection transformation was used to correct the posture of the apple tilted in posture.The results show that the relative error between this method and the actual apple volume is less than 2%,the absolute maximum absolute error is not more than 6 cm^(3),and the detection accuracy is high.When the vertical inclination angle of apples reaches 15°,the absolute absolute error between the method and the actual volume is greater than 6 cm^(3),and the method fails.The error after posture correction is less than 6 cm^(3),which proves that posture correction meets the volume measured by this method.The method proposed in this paper can greatly improve the accuracy of apple volume measurement and provide a favorable technical reference for apple grading.