基于三维点云的采后香蕉表征褐变定量评估方法

    A quantitative evaluation method for postharvest banana characteristic browning based on the three-dimensional point cloud

    • 摘要:
      目的 研究采后香蕉的表征褐变并评估其衰老程度对香蕉保鲜管理至关重要,本研究致力于解决传统人工测量香蕉表征褐变存在的劳动强度大、效率低下的问题。
      方法 提出一种基于三维点云的采后香蕉表征褐变过程定量评估方法。首先利用三维扫描仪获取香蕉的三维点云模型,重构出香蕉的几何模型;然后使用欧式聚类对香蕉几何模型进行点云滤波降噪处理;再结合图像阈值分割法与散点轮廓算法(Alpha Shapes)求出香蕉的体积、表面积和黑斑面积;最后利用傅里叶函数对香蕉表面黑斑变化过程进行模拟,确定香蕉表征褐变过程的评估模型。设计本算法与溢水法测量实际香蕉体积、手绘测量面积的对比试验。
      结果 拟合香蕉的生长函数,回归直线对观测值的拟合程度R2 =0.9816>0.75,验证了算法的有效性。对比试验结果表明,本算法与实际测量值的平均相对误差小于1%,验证了该算法的准确性和可行性。
      结论 本研究可为香蕉的保鲜管理提供数据及技术支撑。

       

      Abstract:
      Objective To study the characteristic browning of postharvest bananas and evaluate the degree of senescence is very important for their fresh-keeping management. This study aims to solve the problems of high labor intensity and low efficiency of traditional methods for measuring the characteristic browning of bananas.
      Method A quantitative evaluation method of postharvest banana characteristic browning process based on three-dimensional point cloud was proposed. Firstly, three-dimensional point cloud model of banana was obtained by 3D scanner, and the geometric model of banana was reconstructed. Then the banana geometry model was denoised by point cloud filtering using euclidean clustering. The volume, surface area and dark spot area of banana were obtained by combining the image threshold segmentation method and Alpha Shapes. Finally, Fourier function was used to simulate the changing process of banana surface dark spots and determine the evaluation model of banana characteristic browning process. A comparison test was designed between the algorithm and the overflow method to measure the actual banana volume and the hand-drawn measurement area.
      Result By fitting the growth functions of different bananas, the fitting degree R2 of the regression line to the observed values was 0.9816>0.75, indicating the effectiveness of the algorithm. The comparison results showed that the average relative error between the proposed algorithm and the actual measured value was less than 1%, which verified the accuracy and feasibility of the proposed algorithm.
      Conclusion This study can provide data and technical support for banana preservation management.

       

    /

    返回文章
    返回