A new method for pedicel/peduncle detection and size assessment of grapevine berries and other fruits by image analysis
A new method for pedicel/peduncle detection and size assessment of grapevine berries and other fruits by image analysis
复制标题
DOI：
10.1016/j.biosystemseng.2013.06.007
复制
发表时间：
2014-01-01
影响因子：
5.1
通讯作者：
Aleixos, Nuria
中科院分区：
文献类型：
--
作者：
Cubero, Sergio;Diago, Maria Paz;Aleixos, Nuria
酿酒葡萄的浆果大小通常被认为会影响葡萄酒的成分和质量，因为它与浆果的果皮与果肉的比例以及对葡萄酒质量起着关键作用的果皮化合物的浓度有关。酿酒葡萄的大小和重量通常是手工测量的，这是一个缓慢、乏味且不准确的过程。本文重点关注两个主要目标，旨在利用图像分析实现这一过程的自动化：（1）开发一种快速准确的方法来检测和去除浆果图像中的花梗，以及（2）准确确定浆果的大小和重量。提出了一种基于新颖的轮廓特征来检测水果花梗的方法。该方法是专门针对葡萄浆果开发的，后来扩展到一组独立的不同形状和大小的其他水果（例如辣椒、梨、苹果或柑橘）并进行了测试。使用这种方法，该系统能够正确估计酿酒葡萄的浆果重量（R-2 > 0.96）和大小（R-2 > 0.97），并评估其他水果的大小，如柑橘、苹果、梨和红辣椒（R-2 > 0.93）。所开发的图像分析方法的性能经过验证，可以轻松地在自动检测系统中实施，以准确估计包括酿酒葡萄在内的各种水果的重量。在这种情况下，在去梗后的分选台上实施该系统可以为酿酒师提供有关葡萄酒潜在质量的非常有用的信息。 (C) 2013 年 IAgrE。由爱思唯尔有限公司出版。保留所有权利。
The berry size of wine-grapes has often been considered to influence wine composition and quality, as it is related to the skin-to-pulp ratio of the berry and the concentration of skin-located compounds that play a key role in the wine quality. The size and weight of wine-grapes are usually measured by hand, making it a slow, tedious and inaccurate process. This paper focuses on two main objectives aimed at automating this process using image analysis: (1) to develop a fast and accurate method for detecting and removing the pedicel in images of berries, and (2) to accurately determine the size and weight of the berry. A method to detect the peduncle of fruits is presented based on a novel signature of the contour. This method has been developed specifically for grapevine berries, and was later extended and tested with an independent set of other fruits with different shapes and sizes such as peppers, pears, apples or mandarins. Using this approach, the system has been capable of correctly estimating the berry weight (R-2 > 0.96) and size (R-2 > 0.97) of wine-grapes and of assessing the size of other fruits like mandarins, apples, pears and red peppers (R-2 > 0.93). The proven performance of the image analysis methodology developed may be easily implemented in automated inspection systems to accurately estimate the weight of a wide range of fruits including wine-grapes. In this case, the implementation of this system on sorting tables after de-stemming may provide the winemaker with very useful information about the potential quality of the wine. (C) 2013 IAgrE. Published by Elsevier Ltd. All rights reserved.