Monocular-vision measurement of street-tree parameters by integrating depth-space correction and geometric anchors

TANG Fang; CHEN Maolin; XUE Mei; TIAN Siyi; JI Cuicui

您当前的位置：

首页 >

文章列表页 >

Monocular-vision measurement of street-tree parameters by integrating depth-space correction and geometric anchors

Optical Sensing and Measurement | 更新时间：2026-07-23

- Monocular-vision measurement of street-tree parameters by integrating depth-space correction and geometric anchors
- OPTICAL TECHNIQUE Vol. 52, Issue 4, Pages: 459-465(2026)
- 作者机构：
  
  1.重庆交通大学智慧城市学院，重庆 400074
  2.自然资源部地理国情监测重点实验室，重庆 400020
  3.重庆市测绘科学技术研究院，重庆 400020
  4.武汉市测绘研究院，湖北武汉430022
- 作者简介：
- 基金信息：
- DOI：
  CLC： P237;TB811
- Received：01 April 2026，
  
  Revised：2026-05-22，
  
  Accepted：12 June 2026，
  
  Online First：23 July 2026，
  
  Published：20 July 2026
- 稿件说明：
移动端阅览
唐芳,陈茂霖,薛梅等.融合深度空间校正与几何锚点的单目视觉行道树参数测量[J].光学技术,2026,52(04):459-465.

TANG Fang,CHEN Maolin,XUE Mei,et al.Monocular-vision measurement of street-tree parameters by integrating depth-space correction and geometric anchors[J].OPTICAL TECHNIQUE,2026,52(04):459-465.
唐芳,陈茂霖,薛梅等.融合深度空间校正与几何锚点的单目视觉行道树参数测量[J].光学技术,2026,52(04):459-465. DOI：

TANG Fang,CHEN Maolin,XUE Mei,et al.Monocular-vision measurement of street-tree parameters by integrating depth-space correction and geometric anchors[J].OPTICAL TECHNIQUE,2026,52(04):459-465. DOI：

摘要

为实现低成本的行道树结构参数自动提取，提出一种融合深度信息与几何约束的单幅影像测量方法。首先利用实例分割与单目深度估计获取树木区域及相对深度；随后通过视差倒数变换提升深度的几何可用性，并结合几何约束与鲁棒统计实现尺度恢复；最后基于重建的度量深度计算树高、胸径与冠幅。在43株样本上的实验表明，相较线性归一化基线，方法有效提升了测量精度：树高MAE为0.70m（基线3.80m）、胸径MAE为7.19cm（基线21.51cm）、冠幅MAE为0.66m（基线2.35m），误差整体下降。方法在单幅影像条件下即可获得稳定可靠的结构参数，为城市林业调查与碳汇评估提供了高效、低成本方案。

Abstract

To enable low-cost automatic extraction of street tree structural parameters

A single-image measurement method that integrates depth information with geometric constraints is proposed. First

instance segmentation and monocular depth estimation are employed to obtain tree regions and relative depth. Then

an inverse disparity transformation is applied to enhance the geometric usability of depth

and geometric constraints combined with robust statistics are introduced to achieve scale recovery. Finally

tree height

diameter at breast height （DBH）

and crown width are estimated based on the reconstructed metric depth.Experiments on 43 samples demonstrate that

compared with a linear normalization baseline

the proposed method significantly improves measurement accuracy: the MAE of tree height is 0.70 m （baseline: 3.80 m）

DBH is 7.19 cm （baseline: 21.51 cm）

and crown width is 0.66 m （baseline: 2.35 m）

with overall error reduction. The method can obtain stable and reliable structural parameters from a single image

providing an efficient and low-cost solution for urban forestry surveys and carbon stock assessment.

关键词

Keywords

references

梁陈涛 , 杨艳波 , 田盼立 , 等 . 基于街景测量的南昌市行道树结构特征与健康状况研究 [J ] . 生态学报 , 2022 , 42 （ 02 ）: 549 — 560 .

Liang Chentao , Yang Yanbo , Tian Panli , et al . Structural characteristics and health status of street trees in Nanchang based on street view measurement [J ] . Acta Ecologica Sinica , 2022 , 42 （ 2 ）: 549 — 560 .

曹明兰 , 张力小 , 王强 . 无人机遥感影像中行道树信息快速提取 [J ] . 中南林业科技大学学报 , 2016 , 36 （ 10 ）: 89 — 93 .

Cao Minglan , Zhang Lixiao , Wang Qiang . Rapid extraction of street tree information from UAV remote sensing images [J ] . Journal of Central South University of Forestry & Technology , 2016 , 36 （ 10 ）: 89 — 93 .

Anderson K , Hancock S , Casalegno S , et al . Visualizing the urban green volume: Exploring LiDAR voxels with tangible technologies and virtual models [J ] . Landscape and Urban Planning , 2018 , 178 : 248 - 260 .

刘也 , 梁文彪 . 车载激光点云在城市行道树普查几何属性提取中的应用 [J ] . 北京测绘 , 2024 , 38 （ 11 ）: 1599 — 1603 .

Liu Ye , Liang Wenbiao . Application of vehicle-borne laser point cloud in geometric attribute extraction for urban street tree census [J ] . Beijing Surveying and Mapping , 2024 , 38 （ 11 ）: 1599 — 1603 .

郭泽宇 , 唐丽玉 , 陈季委 . 基于手持激光点云数据的园林树木参数关系模型 [J ] . 应用激光 , 2025 , 45 （ 5 ）: 138 — 147 .

Guo Zeyu , Tang Liyu , Chen Jiwei . Parameter relationship model of garden trees based on handheld LiDAR point cloud data [J ] . Applied Laser , 2025 , 45 （ 5 ）: 138 — 147 .

崔庆华 , 梁照东 , 张兴安 , 等 . 融合无人机和背包激光雷达数据的森林地上生物量快速估算研究 [J ] . 山东林业科技 , 2025 , 55 （ 2 ）: 9 — 17 .

Cui Qinghua , Liang Zhaodong , Zhang Xing'An , et al . Rapid estimation of forest aboveground biomass by integrating UAV and backpack LiDAR data [J ] . Shandong Forestry Science and Technology , 2025 , 55 （ 2 ）: 9 - 17 .

王怡 , 胡文杰 , 孙拥康 , 等 . 基于近景摄影测量技术的混交林空间结构分析 [J ] . 森林与环境学报 , 2025 , 45 （ 06 ）: 629 — 640 .

Wang Yi , Hu Wenjie , Sun Yongkang , et al . Spatial structure analysis of mixed forests based on close-range photogrammetry [J ] . Journal of Forest and Environment , 2025 , 45 （ 6 ）: 629 - 640 .

尹萍 , 徐爱俊 , 叶俊华 , 等 . 基于消费级双目相机的立木因子测量方法 [J ] . 浙江农林大学学报 , 2023 , 40 （ 02 ）: 436 — 445 .

Yin Ping , Xu Aijun , Ye Junhua , et al . Measuring method of standing tree factors based on consumer-grade binocular camera [J ] . Journal of Zhejiang A&F University , 2023 , 40 （ 2 ）: 436 — 445 .

Khan D , Zheng Z , Uchiyama H , et al . Recent advances in vision-based indoor navigation: A systematic literature review [J ] . Computers & Graphics , 2022 , 104 : 24 — 45 . DOI: 10.1016/j.cag.2022.03.005 http://dx.doi.org/10.1016/j.cag.2022.03.005 .

侯鑫新 , 谭月胜 , 钱桦 , 等 . 一种基于单CCD与经纬仪的树木胸径测量方法 [J ] . 计算机应用研究 , 2014 , 31 （ 4 ）: 1225 — 1228 .

Hou Xinxin , Tan Yuesheng , Qian Hua , et al . A method for measuring tree diameter at breast height based on a single CCD and theodolite [J ] . Computer Application Research , 2014 , 31 （ 4 ）: 1225 — 1228 .

高莉平 , 徐爱俊 . 应用智能终端的立木高度测量方法 [J ] . 东北林业大学学报 , 2018 , 46 （ 11 ）: 28 — 34 .

Gao Liping , Xu Aijun . Tree height measurement method using intelligent terminals [J ] . Journal of Northeast Forestry University , 2018 , 46 （ 11 ）: 28 — 34 .

Coelho J , Fidalgo B , Crisóstomo M M , et al . Non-destructive fast estimation of tree stem height and volume using image processing [J ] . Symmetry , 2021 , 13 （ 3 ）: 374 .

Wu X , Zhou S , Xu A , et al . Passive measurement method of tree diameter at breast height using a smartphone [J ] . Computers and Electronics in Agriculture , 2019 , 163 : 104875 .

Jia J , Kang J , Chen L , et al . A comprehensive evaluation of monocular depth estimation methods in low-altitude forest environment [J ] . Remote Sensing , 2025 , 17 （ 4 ）: 717 .

Rege Cambrin D , Corley I , Garza P . Depth any canopy: Leveraging depth foundation models for canopy height estimation [C ] // European Conference on Computer Vision . Cham : Springer Nature Switzerland , 2024 : 71 — 86 .

Zhao G , Cai W , Wang Z , et al . Phenotypic parameters estimation of plants using deep learning-based 3-D reconstruction from single RGB image [J ] . IEEE Geoscience and Remote Sensing Letters , 2022 , 19 : 1 — 5 .

赵永杰 , 蒲六如 , 宋磊 , 等 . 基于Shuffle-ZoeDepth单目深度估计的苗期玉米株高测量方法 [J ] . 农业机械学报 , 2024 , 55 （ 5 ）: 222 — 230 .

Zhao Yongjie , Pu Liuru , Song Lei , et al . Measurement method of seedling stage maize height based on Shuffle-ZoeDepth monocular depth estimation [J ] . Transactions of the Chinese Society for Agricultural Machinery , 2024 , 55 （ 5 ）: 222 — 230 .

Yang L , Kang B , Huang Z , et al . Depth anything v2 [J ] . Advances in Neural Information Processing Systems , 2024 , 37 : 21875 — 21911 .

Ravi N , Gabeur V , Hu Y T , et al . Sam 2: Segment anything in images and videos [C ] // International Conference on Learning Representations . Vienna, Austria : ICLR , 2025 : 28085 — 28128 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Study on a method for evaluating geometric tolerances based on monocular rotational vision

Development of a high-resolution white-light interference and confocal microscopic sensor for three-dimensional surface topography measurement

Research review on the detection of trace gases based on DOAS technology

Square-wave modulation thermoreflectance for metal film thickness measurement

Research on Shared Management and Operation Mechanism of University Public Experimental Platforms

Related Author

No data

Related Institution

No data

⁰