基于数字轮廓图像的水果体积估计

Volume Estimation of Fruit from Digital Image

在分析比较各种可能的技术方案的基础上,本文提出基于水果轮廓特征的人工神经网络估计方法,并建立了特征库。实验结果表明,该方法优于Miller和Marchant的方法,并且不同的特征对不同水果的体积估计的影响是不同的。

Based on the analysis and comparison of the possible solutions to the volume estimation, an artificial neural network method based on the profile shape features is proposed and a corresponding feature pool is established. The experimental results show that this method is better than the solutions proposed by Miller and Marchant, and that the contributions of different features vary depending on the fruits.