Cotton plant point cloud completion by collaborative segmentation and improved completion network

doi:10.1016/j.jia.2025.10.006

Advanced Online Publication | Current Issue | Archive | Adv Search

Chunjing Si^{1, 3}, Zhiben Yin⁴, Liping Chen¹, Xiangyang Li⁵, Mingdeng Shi^1,3#, Xuping Feng², Tiecheng Bai^{1, 3#}, Yong He²

¹College of Information Engineering, Tarim University, Alaer 843300, China

²College of Biosystems Engineering and Food Science, Zhejiang University, Hangzhou 310058, China

³Key Laboratory of Tarim Oasis Agriculture (Tarim University), Ministry of Education, Alaer 843300, China

⁴University of Information Science and Engineering, Xinjiang University of Science & Technology, Korla 841000, China

⁵Tarim University Library, Tarim University, Alaer 843300, China

Highlights

Ÿ Establish a Cotton 3D dataset comprising over 724 high-quality cotton plants.

Ÿ Complete the morphology of cotton leaves from the whole cotton plant based on semantic segmentation.

Ÿ Enhance PF-Net through a unified loss function to improve the quality of cotton leaf morphological features.

Abstract
References

Download: PDF in ScienceDirect
Export: BibTeX | EndNote (RIS)

摘要

实现棉花叶片形态点云模型的精准补全，对于系统研究叶片形态参数与环境条件间的互作规律具有重要意义。现有方法在处理形态结构复杂多样的植物三维缺失点云模型时存在局限性。本研究提出一种面向棉花叶片点云补全的PCompNet模型。该模型通过融合基于深度层次化特征学习的形态部件分割技术，可从形态多样、表面不连续的整株棉花中实现点云叶片几何形态的完整补全。此外，引入均匀化损失函数，能有效惩罚 PF-Net网络中面片中心与其最近邻点的平均距离偏差，进而避免生成的棉花叶片缺失区域点云数据出现过度集中现象。实验结果表明，在Cotton3D数据集上，与 PMP-Net、GRNet、SnowfakeNet、FoldingNet 和 PF-Net 相比，PCompNet 的倒角距离（CD）分别降低了 95.46%、98.45%、97.46%、100.00% 和 84.93%。此外，PCompNet 能够在保留输入点云几何形态的同时，精准补全不同尺度的缺失区域。即使在 75% 数据缺失的情况下，其 CD 值仍可维持在 0.115。上述研究结果验证了 PCompNet 在棉花叶片点云模型补全任务中的有效性与稳健性，同时也揭示了在棉花生长动态监测、环境适应性评估等相关领域的潜在应用前景。

Abstract

Cotton leaves are fundamental components for cotton growth and serve vital roles in photosynthesis and transpiration. The completion of point cloud data on cotton leaf morphology is critically important for examining the interaction between morphological parameters and the environment. Previous methods have shown effective performance in capturing objects with regular shapes and continuous surfaces, particularly for industrially produced 3D-modeled objects. However, these techniques demonstrate limitations in processing plants with diverse morphological structures. This study proposes PCompNet (a segmentation and improved completion network) for cotton leaf point cloud completion, reconstructing complete geometries from whole plants with diverse shapes and discontinuous surfaces through morphological part segmentation technique with deep hierarchical point-set feature learning. Additionally, a unified loss function was implemented to effectively penalize the average distance discrepancy between patch centers and their nearest neighbors in PF-Net, preventing the generated missing point clouds of cotton leaves from excessive concentration. The experimental results demonstrated that PCompNet achieved substantial reductions in Chamfer distance (CD) on the Cotton3D dataset compared to PMP-Net, GRNet, SnowfakeNet, FoldingNet, and PF-Net, with reductions of 95.46, 98.45, 97.46, 100.00, and 84.93%, respectively. Moreover, PCompNet accurately completed missing regions at different scales while maintaining the geometry of the input point cloud. Even with 75% of data missing, the CD value remained at 0.115. These results demonstrate the effectiveness and robustness of PCompNet in completing point cloud data for cotton leaves, indicating its potential for applications in cotton growth and environmental studies.

Keywords: part segmentation unified loss function Cotton3D dataset point cloud completion cotton plant

Online: 21 October 2025

Fund:

This work was supported by the National Natural Science Foundation of China (61961035 and 61961034), and the Xinjiang Production and Construction Corps Science and Technology Project, China (2023AB063).

About author: #Correspondence Tiecheng Bai, E-mail: baitiecheng1983@163.com; Mingdeng Shi, E-mail: smdeng@taru.edu.cn

	Service
	E-mail this article
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors

Cite this article:

Chunjing Si, Zhiben Yin, Liping Chen, Xiangyang Li, Mingdeng Shi, Xuping Feng, Tiecheng Bai, Yong He. 2025. Cotton plant point cloud completion by collaborative segmentation and improved completion network. Journal of Integrative Agriculture, Doi:10.1016/j.jia.2025.10.006

Achlioptas P, Diamanti O, Mitliagkas I, Guibas L. 2018. Learning representations and generative models for 3D point clouds. PMLR, 80, 40-49.

Brown T B, Cheng R, Sirault X R, Rungrat T, Murray K D, Trtilek M, Furbank R T, Badger M, Pogson B J, Borevitz J O. 2014. TraitCapture: Genomic and environment modelling of plant phenomic data. Current Opinion in Plant Biology, 18, 73-79.

Chang Y, Jung C, Xu Y. 2021. FinerPCN: High fidelity point cloud completion network using pointwise convolution. Neurocomputing, 460, 266-276.

Chen H, Liu S, Wang C, Wang C, Gong K, Li Y, Lan Y. 2023. Point cloud completion of plant leaves under occlusion conditions based on deep learning. Plant Phenomics, 5, 0117.

Debnath S, Paul M, Debnath T. 2023. Applications of LiDAR in agriculture and future research directions. Journal of Imaging, 9, 57.

Du R, Ma Z, Xie P, He Y, Cen H. 2023. PST: Plant segmentation transformer for 3D point clouds of rapeseed plants at the podding stage. ISPRS Journal of Photogrammetry and Remote Sensing, 195, 380-392.

Ester M, Kriegel H P, Sander J, Xu X. 1996. A density-based algorithm for discovering clusters in large spatial databases with noise. kdd, 96, 226-231.

Fan H, Su H, Guibas L. 2017. A point set generation network for 3D object reconstruction from a single image. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Hawaii, USA. pp. 605-613.

Fei B, Yang W, Chen W M, Li Z, Li Y, Ma T, Hu X, Ma L. 2022. Comprehensive review of deep learning-based 3D point cloud completion processing and analysis. IEEE Transactions on Intelligent Transportation Systems, 23, 22862-22883.

Goodfellow I, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, Courville A, Bengio Y. 2020. Generative adversarial networks. Communications of the ACM, 63, 139-144.

Han B, Li Y, Bie Z, Peng C, Huang Y, Xu S. 2022. MIX-NET: Deep learning-based point cloud processing method for segmentation and occlusion leaf restoration of seedlings. Plants (Basel), 11, 3342.

Hu C, Li P, Pan Z. 2018. Phenotyping of poplar seedling leaves based on a 3D visualization method. International Journal of Agricultural and Biological Engineering, 11, 145-151.

Huang Z, Yu Y, Xu J, Ni F, Le X. 2020. PF-Net: Point fractal network for 3D point cloud completion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Washington, USA. pp. 7662-7670.

Jin S, Su Y, Gao S, Wu F, Ma Q, Xu K, Ma Q, Hu T, Liu J, Pang S, Guan H, Zhang J, Guo Q. 2020. Separating the structural components of maize for field phenotyping using terrestrial LiDAR data and deep convolutional neural networks. IEEE Transactions on Geoscience and Remote Sensing, 58, 2644-2658.

Li R, Li X, Fu C W, Cohen-Or D, Heng P A. 2019. PU-GAN: A Point cloud upsampling adversarial network. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, Seoul, South Korea. pp. 7203-7212.

Lin C, Hu F, Peng J, Wang J, Zhai R. 2022. Segmentation and stratification methods of field maize terrestrial LiDAR point cloud. Agriculture, 12, 1450.

Liu M, Sheng L, Yang S, Shao J, Hu S M. 2019. Morphing and sampling network for dense point cloud completion. In: Proceedings of the AAAI Conference on Artificial Intelligence, Hawaii, USA. pp. 11596-11603.

Liu Y, Wu X Q. 2023. 3D shape completion via deep learning: A method survey. Journal of Graphics, 44, 201-215.

Lu H, Shi H. 2020. Deep learning for 3D point cloud understanding: A survey. arXiv preprint, arXiv:2009.08920.

Luo L, Jiang X, Yang Y, Samy E R A, Lefsrud M, Hoyos-Villegas V, Sun S. 2023. Eff-3dpseg: 3d organ-level plant shoot segmentation using annotation-efficient deep learning. Plant phenomics, 5, 0080.

Ma X, Zhu K, Guan H, Feng J, Yu S, Liu G. 2019. High-throughput phenotyping analysis of potted soybean plants using colorized depth images based on a proximal platform. Remote Sensing, 11, 1085.

Nguyen D T, Hua B S, Tran M K, Pham Q H, Yeung S K. 2016. A field model for repairing 3D shapes. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Nevada, USA. pp. 5676–5684.

Qi C R, Yi L, Su H, Guibas L J. 2017. PointNet++: Deep hierarchical feature learning on point sets in a metric space. Advances in Neural Information Processing Systems, 30, 1-10.

Schunck D, Magistri F, Rosu R A, Cornelissen A, Chebrolu N, Paulus S, Leon J, Behnke S, Stachniss C, Kuhlmann H, Klingbeil L. 2021. Pheno4D: A spatio-temporal dataset of maize and tomato plant point clouds for phenotyping and advanced plant analysis. PLoS ONE, 16, e0256340.

Thakur S, Bawiskar S, Singh S K, Shanmugasundaram M. 2020. Autonomous farming-visualization of image processing in agriculture. In Proceedings of Inventive Communication and Computational Technologies, Perambalur, India. pp. 345-351.

Ullman S. 1979. The interpretation of structure from motion. Proceedings of the Royal Society of London, 203, 405-426.

Wang Y, Chen Y, Zhang X, Gong W. 2021. Research on measurement method of leaf length and width based on point cloud. Agriculture, 11, 63.

Wen X, Xiang P, Han Z, Cao Y P, Wan P, Zheng W, Liu Y S. 2021. PMP-Net: Point cloud completion by learning multi-step point moving paths. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Tennessee, USA. pp. 7443-7452.

Wen X, Xiang P, Han Z, Cao Y P, Wan P, Zheng W, Liu Y S. 2023. PMP-Net++: Point cloud completion by transformer-enhanced multi-step point moving paths. IEEE Trans Pattern Anal Mach Intell, 45, 852-867.

Xia T, Sheng W, Song R, Li H, Zhang M. 2023. A review of three-dimensional multispectral imaging in plant phenotyping. In: Sensing Technologies for Field and In-House Crop Production (Technology Review and Case Studies). Springer. pp. 1-18.

Xiang P, Wen X, Liu Y S, Cao Y P, Wan P, Zheng W, Han Z. 2021. SnowflakeNet: Point cloud completion by snowflake point deconvolution with skip-transformer. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, Quebec, Canada. pp. 5499-5509.

Xie H, Yao H, Zhou S, Mao J, Zhang S, Sun W. 2020. GRNet: Gridding residual network for dense point cloud completion. In: Proceedings of the European Conference on Computer Vision, Scotland, United Kingdom. pp. 365-381.

Xue J, Hou X, Zeng Y. 2021. Review of image-based 3D reconstruction of building for automated construction progress monitoring. Applied Sciences, 11, 7840.

Yang R, He Y, Lu X, Zhao Y, Li Y, Yang Y, Kong W, Liu F. 2024. 3D-based precise evaluation pipeline for maize ear rot using multi-view stereo reconstruction and point cloud semantic segmentation. Computers and Electronics in Agriculture, 216, 108512.

Yang Y, Feng C, Shen Y, Tian D. 2017. FoldingNet: Point cloud auto-encoder via deep grid deformation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Hawaii, USA. pp. 206-215.

Yuan W, Khot T, Held D, Mertz C, Hebert M. 2018. PCN: Point completion network. In: Proceeding of the 2018 International Conference on 3D Vision (3DV), Verona, Italy. pp. 728-737.

Yusri A, Haziq M, Ma J, Ns K, Mhm R. 2022. Preservation of cultural heritage: A comparison study of 3D modelling between laser scanning, depth image, and photogrammetry methods. Journal of Mechanical Engineering (JMechE), 19, 125-146.

Zhang J, Chen X, Cai Z, Pan L. 2021. Unsupervised 3D shape completion through GAN inversion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Tennessee, USA. pp. 1768-1777.

Zhou J, Wu Y, Chen J, Cui M, Gao Y, Meng K, Wu M, Guo X, Wen W. 2023. Maize stem contour extraction and diameter measurement based on adaptive threshold segmentation in field conditions. Agriculture, 13, 678.

[1]	Niu Wang, Weidong Zhang, Zhenyu Zhong, Xiongbo Zhou, Xinran Shi, Xin Wang. FGF7 secreted from dermal papillae cell regulates the proliferation and differentiation of hair follicle stem cell[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3583-3597.
[2]	Lichao Zhai, Shijia Song, Lihua Zhang, Jinan Huang, Lihua Lv, Zhiqiang Dong, Yongzeng Cui, Mengjing Zheng, Wanbin Hou, Jingting Zhang, Yanrong Yao, Yanhong Cui, Xiuling Jia. Subsoiling before winter wheat alleviates the kernel position effect of densely grown summer maize by delaying post-silking root–shoot senescence[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3384-3402.
[3]	Tiantian Chen, Lei Li, Dan Liu, Yubing Tian, Lingli Li, Jianqi Zeng, Awais Rasheed, Shuanghe Cao, Xianchun Xia, Zhonghu He, Jindong Liu, Yong Zhang. Genome wide linkage mapping for black point resistance in a recombinant inbred line population of Zhongmai 578 and Jimai 22[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3311-3321.
[4]	Zuxian Chen, Bingbing Zhao, Yingying Wang, Yuqing Du, Siyu Feng, Junsheng Zhang, Luxiang Zhao, Weiqiang Li, Yangbao Ding, Peirong Jiao. H5N1 avian influenza virus PB2 antagonizes duck IFN-β signaling pathway by targeting mitochondrial antiviral signaling protein[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3614-3625.
[5]	Yang Sun, Yu Liu, Li Zhou, Xinyan Liu, Kun Wang, Xing Chen, Chuanqing Zhang, Yu Chen. Activity of fungicide cyclobutrifluram against Fusarium fujikuroi and mechanism of the pathogen resistance associated with point mutations in FfSdhB, FfSdhC₂ and FfSdhD[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3511-3528.
[6]	Yufeng Xiao, Meiqi Dong, Xian Wu, Shuang Liang, Ranhong Li, Hongyu Pan, Hao Zhang. Enrichment, domestication, degradation, adaptive mechanism, and nicosulfuron bioremediation of bacteria consortium YM2[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3529-3545.
[7]	Yuxin He, Fei Deng, Chi Zhang, Qiuping Li, Xiaofan Huang, Chenyan He, Xiaofeng Ai, Yujie Yuan, Li Wang, Hong Cheng, Tao Wang, Youfeng Tao. Wei Zhou, Xiaolong Lei, Yong Chen, Wanjun Ren. Can a delayed sowing date improve the eating and cooking quality of mechanically transplanted rice in the Sichuan Basin, China?[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3368-3383.
[8]	Dili Lai, Md. Nurul Huda, Yawen Xiao, Tanzim Jahan, Wei Li, Yuqi He, Kaixuan Zhang, Jianping Cheng, Jingjun Ruan, Meiliang Zhou. *Evolutionary and expression analysis of sugar transporters from Tartary buckwheat revealed the potential function of FtERD23* in drought stress**[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3334-3350.
[9]	Zishuai Wang, Wangchang Li, Zhonglin Tang. Enhancing the genomic prediction accuracy of swine agricultural economic traits using an expanded one-hot encoding in CNN models[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3574-3582.
[10]	Yunji Xu, Xuelian Weng, Shupeng Tang, Weiyang Zhang, Kuanyu Zhu, Guanglong Zhu, Hao Zhang, Zhiqin Wang, Jianchang Yang. Untargeted lipidomic analysis of milled rice under different alternate wetting and soil drying irrigation regimes[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3351-3367.
[11]	Minghui Li, Yilan Chen, Siqiao Wang, Xueke Sun, Yongkun Du, Siyuan Liu, Ruiqi Li, Zejie Chang, Peiyang Ding, Gaiping Zhang. Plug-and-display nanoparticle immunization of the core epitope domain induces potent neutralizing antibody and cellular immune responses against PEDV[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3598-3613.
[12]	Jing Zhou, Bingshuai Du, Yibo Cao, Kui Liu, Zhihua Ye, Yiming Huang, Lingyun Zhang. Genome-wide identification of sucrose transporter genes in Camellia oleifera and characterization of CoSUT4[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3494-3510.
[13]	Yuheng Wang, Furong Kang, Bo Yu, Quan Long, Huaye Xiong, Jiawei Xie, Dong Li, Xiaojun Shi, Prakash Lakshmanan, Yueqiang Zhang, Fusuo Zhang. Magnesium supply is vital for improving fruit yield, fruit quality and magnesium balance in citrus orchards with increasingly acidic soil[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3641-3655.
[14]	Mingxin Feng, Ying Hu, Xin Yang, Jingwen Li, Haochen Wang, Yujia Liu, Haijun Ma, Kai Li, Jiayin Shang, Yulin Fang, Jiangfei Meng. Uncovering the miRNA-mediated regulatory network involved in postharvest senescence of grape berries[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3465-3483.
[15]	Li Liu, Yifeng Feng, Ziqi Han, Yaxiao Song, Jianhua Guo, Jing Yu, Zidun Wang, Hui Wang, Hua Gao, Yazhou Yang, Yuanji Wang, Zhengyang Zhao. *Functional analysis of the xyloglucan endotransglycosylase/hydrolase gene MdXTH2* in apple fruit firmness formation**[J]. >Journal of Integrative Agriculture, 2025, 24(9): 3418-3434.

No Suggested Reading articles found!

Viewed

Full text

Abstract

Cited

Shared

Discussed

Cite this article:

About JIA

Editorial board

For authors