[1]魏超宇,韩文,庞程,等.基于多尺度特征融合和密集连接网络的疏果期黄花梨植株图像分割[J].江苏农业学报,2021,(04):990-997.[doi:doi:10.3969/j.issn.1000-4440.2021.04.023]
　WEI Chao-yu,HAN Wen,PANG Cheng,et al.Image segmentation of Huanghua pear plants at fruit-thinning stage based on multi-scale feature fusion and dense connection network[J].,2021,(04):990-997.[doi:doi:10.3969/j.issn.1000-4440.2021.04.023]
点击复制

基于多尺度特征融合和密集连接网络的疏果期黄花梨植株图像分割()
分享到：

江苏农业学报[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2021年04期
页码:
990-997
栏目:
园艺
出版日期:
2021-08-28

文章信息/Info

Title:
Image segmentation of Huanghua pear plants at fruit-thinning stage based on multi-scale feature fusion and dense connection network
作者:
魏超宇; 韩文; 庞程; 刘辉军
（中国计量大学计量测试工程学院，浙江杭州310018）
Author(s):
WEI Chao-yu; HAN Wen; PANG Cheng; LIU Hui-jun
(College of Metrological Technology and Engineering, China Jiliang University, Hangzhou 310018, China)
关键词:
黄花梨植株; 多尺度特征融合; 密集连接网络; 图像分割; 空洞空间金字塔池化（ASPP）; 感受野
Keywords:
Huanghua pear plants; multi-scale feature fusion; dense connection network; image segmentation; atrous spatial pyramid pooling (ASPP); receptive field
分类号:
TP391.4
DOI:
doi:10.3969/j.issn.1000-4440.2021.04.023
文献标志码:
A
摘要:
由于自然环境下果蔬植株的果实、枝干和叶片等目标尺度不一、边缘不规则，因此造成其准确分割较为困难。针对该问题，提出1种多尺度特征融合和密集连接网络（Multi-scale feature fusion and dense connection networks，MDNet）以实现黄花梨疏果期植株图像的准确分割。在研究中借鉴了编码-解码网络，其中编码网络采用DenseNet对多层特征进行复用和融合，以改善信息传递方式；解码网络使用转置卷积进行上采样，结合跳层连接融合浅层细节信息与深层语义信息；在编码、解码之间加入空洞空间金字塔池化（Atrous spatial pyramid pooling，ASPP）用于提取不同感受野的特征图以融合多尺度特征，聚合上下文信息。结果表明，ASPP有效提高了模型的分割精度，MDNet在测试集上的平均局域重合度（MIoU）为77.97%，分别较SegNet、Deeplabv2和DNet提高了8.10个、5.77个和2.17个百分点，果实、枝干和叶片的像素准确率分别为93.57%、90.31%和95.43%，实现了黄花梨植株果实、枝干和叶片等目标的准确分割。在翠冠梨植株图像的独立测试中，MIoU为70.93%，表明该模型具有较强的泛化能力，对自然环境下果蔬植株图像的分割有一定的参考价值。
Abstract:
As the fruits, branches and leaves of fruit and vegetable plants vary in scales and margins under natural environment, it is difficult to segment them accurately. To solve the problem, a multi-scale feature fusion and dense connection network (MDNet) was proposed to achieve the accurate segmentation of Huanghua pear images at fruit-thinning stage. The coding-decoding network was adopted in this study and DenseNet was adopted to reuse and fuse multi-layer features in the coding network, so as to improve the transfer mode of information. The transposed convolution was used to carry out up-sampling in the decoding network, and the skip-layer connection was also employed to fuse shallow detail information and deep semantic information. The atrous spatial pyramid pooling (ASPP) was added between coding and decoding to extract future maps with different receptive fields so as to fuse multi-scale feature and aggregate context information. The results showed that ASPP improved the segmentation accuracy of the MDNet model effectively. The mean intersection over union (MIoU) of the MDNet on the test set was 77.97%, improved by 8.10, 5.77 and 2.17 percentage points respectively compared with SegNet, Deeplabv2 and DNet. The pixel accuracy for fruits, branches and leaves was 93.57%, 90.31% and 95.43%, respectively. Therefore, the accurate segmentation of fruits, branches and leaves of Huanghua pear plants was realized. The MIoU was 70.93% in the independent test of Cuiguan pear plants images, indicating that the model had strong generalization ability and was particularly valuable for the image segmentation of fruit and vegetable plants in natural environment.

参考文献/References:

[1]GUO W, RAGE U K, NINOMIYA S. Illumination invariant segmentation of vegetation for time series wheat images based on decision tree model[J]. Computers and Electronics in Agriculture, 2013, 96: 58-66.
[2]XIONG X, DUAN L F, LIU L B, et al. Panicle-SEG: a robust image segmentation method for rice panicles in the field based on deep learning and superpixel optimization[J]. Plant Methods, 2017, 13(1): 104.
[3]苏伟，蒋坤萍，郭浩，等. 地基激光雷达提取大田玉米植株表型信息[J].农业工程学报，2019，35(10): 125-130.
[4]CHENG H, DAMEROW L, BLANKE M, et al. ANN model for apple yield estimation based on feature of tree image[J]. Transactions of the Chinese Society for Agricultural Machinery, 2015, 46(1): 14-19.
[5]赵德安，刘晓洋，陈玉，等. 苹果采摘机器人夜间识别方法[J].农业机械学报，2015，46(3): 15-22.
[6]ZHANG S W, WANG H X, HUANG W Z, et al. Plant diseased leaf segmentation and recognition by fusion of superpixel, K-means and PHOG[J]. Optik, 2017, 157: 866-872.
[7]ZEMMOUR E, KURTSER P, EDAN Y. Automatic parameter tuning for adaptive thresholding in fruit detection[J]. Sensors, 2019, 19(9), 2130-2151.
[8]卢宏涛，张秦川. 深度卷积神经网络在计算机视觉中的应用研究综述[J].数据采集与处理，2016，31(1): 1-17.
[9]王丹丹，何东健. 基于R-FCN深度卷积神经网络的机器人疏果前苹果目标的识别[J].农业工程学报，2019，35(3): 156-163.
[10]薛月菊，黄宁，涂淑琴，等. 未成熟芒果的改进YOLOv2识别方法[J].农业工程学报，2018，34(7): 173-179.
[11]魏云超，赵 耀. 基于DCNN的图像语义分割综述[J].北京交通大学学报（自然科学版），2016，40(4): 82-91.
[12]LONG J, SHELHAMER, DARRELL T. Fully convolutional networks for semantic segmentation[C]//IEEE. IEEE conference on computer vision and pattern recognition. Seoul, Korea: IEEE, 2015: 3431-3440.
[13]BADRINARAYANAN V, KENDALL A, CIPOLLA R. SegNet: a deep convolutional encoder-decoder architecture for image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39: 2481-2495.
[14]CHEN L C, PAPANDREOU G, KOKKINOS I, et al. DeepLab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40(4): 834-848.
[15]LIN G C, TANG Y C, ZOU X J, et al. Guava detection and pose estimation using a low-cost RGB-D sensor in the field[J]. Sensors, 2019, 19(2): 428-443.
[16]MAJEED Y, ZHANG J, ZHANG X, et al. Apple tree trunk and branch segmentation for automatic trellis training using convolutional neural network based semantic segmentation[J]. International Federation of Automatic Control, 2018, 51(17): 75-80.
[17]AMBROZIO D P, AMY T, HENRY M. Multi-species fruit flower detection using a refined semantic segmentation network[J]. IEEE Robotics and Automation Letters, 2018, 3(4): 3003-3010.
[18]KANG H W, CHEN C. Fruit detection, segmentation and 3D visualisation of environments in apple orchards[J]. Computers and Electronics Agriculture, 2020, 171: 1016-1033.
[19]NOH H, HONG, HAN B. Learning deconvolution network for semantic segmentation[C]//IEEE. IEEE international conference on computer vision. Los Alamitos, USA: IEEE Computer Society Press, 2015: 1520-1528.
[20]HUANG G, LIU Z, WEINBERGER K Q, et al. Densely connected convolutional networks[C]//IEEE. IEEE conference on computer vision and pattern recognition. Honolulu, USA: IEEE, 2016: 4700-4708.
[21]CHEN L C, PAPANDREOU G, KOKKINOS I, et al. Semantic image segmentation with deep convolutional nets and fully connected crfs[J]. International Conference on Learning Representations, 2014(4):357-361.
[22]徐子豪，黄伟泉，王胤. 基于深度学习的监控视频中多类别车辆检测[J].计算机应用，2019，39(3):700-705.
[23]YANG M K, YU K, ZHANG C, et al. DenseASPP for semantic segmentation in street scenes[C]//IEEE. IEEE computer society conference on computer vision and pattern recognition. Piscataway, NJ: IEEE Press, 2018: 3684-3692.
[24]李云伍，徐俊杰，刘得雄，等. 基于改进空洞卷积神经网络的丘陵山区田间道路场景识别[J].农业工程学报，2019，35(7): 150-159.

备注/Memo

备注/Memo:
收稿日期：2020-10-13基金项目：国家自然科学基金项目(51606181)；国家级大学生创新创业训练计划项目(201910356009)作者简介：魏超宇（1995-），男，浙江嘉兴人，硕士研究生，研究方向为计算机视觉、深度学习等。(E-mail)P1802085258@cjlu.edu.cn通讯作者：刘辉军，(E-mail)liuhj@cjlu.edu.cn
更新日期/Last Update: 2021-09-06