Crop segmentation from images by morphology modeling in the CIE L*a*b* color space
Crop segmentation from images by morphology modeling in the CIE L*a*b* color space
复制标题
DOI：
10.1016/j.compag.2013.08.022
复制
发表时间：
2013-11-01
影响因子：
8.3
通讯作者：
Li, C. N.
中科院分区：
文献类型：
--
作者：
Bai, X. D.;Cao, Z. G.;Li, C. N.
从野外拍摄的图像中分割农作物是一项复杂的任务。利用一种新的形态学建模方法，建立了CIE L*a*B* 中的作物颜色模型。(or实验室用于简化）颜色空间，实现农作物图像的分割.在有监督学习阶段，应用形态学建模来处理与像素亮度分量有关的作物颜色特征，并建立作物颜色模型。为了验证该方法的性能，选取了2011年4月27日至2011年5月21日拍摄的56幅大小为601 x 601的测试图像，将该方法与其他8种著名方法进行了比较。实验结果表明，该方法的分割质量约为87.2%的自动目标识别工作组（ATRWG）的评价方法和96.0%的另一种评价方法。此外，对阴天、阴天和晴天的图像分割性能进行了分析。实验表明，该方法对光照变化具有较强的鲁棒性，优于其他8种方法。此外，不同的结构元素类型所提出的方法的影响进行了比较。总体而言，本文提出的农作物分割方法可以有效地应用于田间作物分割。(C)2013爱思唯尔有限公司版权所有。
Crop segmentation from the images taken in the outdoor fields is a complex task. In this paper, a new morphology modeling method is utilized to establish the crop color model in the CIE L*a*b*. (or Lab for simplification) color space and to realize the crop image segmentation. In the supervised learning stage, morphology modeling is applied to deal with the color characteristics of the crop with respect to the pixel lightness component and establish the crop color model. To verify the performance of the proposed method, 56 test images which in size of 601 x 601 and taken from April 27, 2011 to May 21, 2011 are utilized to compare the proposed method with eight other famous approaches. Experiment shows that the segmentation quality of the proposed method is approximately 87.2% for the Automatic Target Recognition Working Group (ATRWG) evaluation method and 96.0% for another evaluation method. Moreover, the segmentation performance for images taken on cloudy, overcast and sunny days is analyzed. Experiment demonstrates that our method is robust to the variation of illumination in the field and performed better than eight other approaches. Furthermore, the impact of different structuring element types to the proposed method is compared. Overall, the proposed crop segmentation method can be used to crop segmentation in the field effectively. (C) 2013 Elsevier B.V. All rights reserved.