Mapping forests using an unmanned ground vehicle with 3D LiDAR and graph-SLAM

Pierzchała, Marek; Giguère, Philippe; Astrup, Rasmus

doi:10.1016/j.compag.2017.12.034

Cited by 190 publications

(143 citation statements)

References 27 publications

Supporting

Mentioning

117

Contrasting

Unclassified

Order By: Relevance

“…By using aerial photogrammetry as a reference, a variety of sensors (such as GPS, IMU, and CCD) have been gradually applied to terrestrial measurement modes. In 2018, Pierzchała et al [34] proposed the method of using a simultaneous localization and mapping (SLAM) algorithm to generate a local forest map and they designed a mobile data acquisition platform that integrates multiple sensors. Lastly, the resulting map was accurately evaluated in the form of a three-dimensional point cloud.…”

Section: Introductionmentioning

confidence: 99%

Extraction of Sample Plot Parameters from 3D Point Cloud Reconstruction Based on Combined RTK and CCD Continuous Photography

et al. 2018

View full text Add to dashboard Cite

Enriching forest resource inventory is important to ensure the sustainable management of forest ecosystems. Obtaining forest inventory data from the field has always been difficult, laborious, time consuming, and expensive. Advances in integrating photogrammetry and computer vision have helped researchers develop some numeric algorithms and methods that can turn 2D (images) into 3D (point clouds) and are highly applicable to forestry. This paper aimed to develop a new, highly accurate methodology that extracts sample plot parameters based on continuous terrestrial photogrammetry. For this purpose, we designed and implemented a terrestrial observation instrument combining real-time kinematic (RTK) and charge-coupled device (CCD) continuous photography. Then, according to the set observation plan, three independent experimental plots were continuously photographed and the 3D point cloud of the plot was generated. From this 3D point cloud, the tree position coordinates, tree DBHs, tree heights, and other plot characteristics of the forest were extracted. The plot characteristics obtained from the 3D point cloud were compared with the measurement data obtained from the field to check the accuracy of our methodology. We obtained the position coordinates of the trees with the positioning accuracy (RMSE) of 0.162 m to 0.201 m. The relative root mean square error (rRMSE) of the trunk diameter measurements was 3.07% to 4.51%, which met the accuracy requirements of traditional forestry surveys. The hypsometrical measurements were due to the occlusion of the forest canopy and the estimated rRMSE was 11.26% to 11.91%, which is still good reference data. Furthermore, these image-based point cloud data also have portable observation instruments, low data collection costs, high field measurement efficiency, automatic data processing, and they can directly extract tree geographic location information, which may be interesting and important for certain applications such as the protection of registered famous trees. For forest inventory, continuous terrestrial photogrammetry with its unique advantages is a solution that deserves future attention in the field of tree detection and ecological construction.

show abstract

Section: Introductionmentioning

confidence: 99%

Extraction of Sample Plot Parameters from 3D Point Cloud Reconstruction Based on Combined RTK and CCD Continuous Photography

et al. 2018

View full text Add to dashboard Cite

show abstract

“…Pierzchala et al [20] used an unmanned vehicle-based MLS system and simultaneous localization and mapping (SLAM) to map trees and estimate diameter at breast height (DBH). A root mean square error (RMSE) of 2.4 cm was obtained when estimating the DBH.…”

Section: Introductionmentioning

confidence: 99%

Comparing Three Different Ground Based Laser Scanning Methods for Tree Stem Detection

et al. 2018

View full text Add to dashboard Cite

Abstract:A forest inventory is often carried out using airborne laser data combined with ground measured reference data. Traditionally, the ground reference data have been collected manually with a caliper combined with land surveying equipment. During recent years, studies have shown that the caliper can be replaced by equipment and methods that capture the ground reference data more efficiently. In this study, we compare three different ground based laser measurement methods: terrestrial laser scanner, handheld laser scanner and a backpack laser scanner. All methods are compared with traditional measurements. The study area is located in southeastern Norway and divided into seven different locations with different terrain morphological characteristics and tree density. The main tree species are boreal, dominated by Norway spruce and Scots pine. To compare the different methods, we analyze the estimated tree stem diameter, tree position and data capture efficiency. The backpack laser scanning method captures the data in one operation. For this method, the estimated diameter at breast height has the smallest mean differences of 0.1 cm, the smallest root mean square error of 2.2 cm and the highest number of detected trees with 87.5%, compared to the handheld laser scanner method and the terrestrial laser scanning method. We conclude that the backpack laser scanner method has the most efficient data capture and can detect the largest number of trees.

show abstract

“…In [4] points are clustered into individual trees; aligned to an existing map, fine tuned by ICP, and diameters at a fixed height extracted. Pierzchala et al [5] improve the performance of offline laser maps in forests with loop closure techniques from SLAM.…”

Section: A Mapping Forestsmentioning

confidence: 99%

Visual Appearance Analysis of Forest Scenes for Monocular SLAM

Garforth

Webb

2019

2019 International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Monocular simultaneous localisation and mapping (SLAM) is a cheap and energy efficient way to enable Unmanned Aerial Vehicles (UAVs) to safely navigate managed forests and gather data crucial for monitoring tree health. SLAM research, however, has mostly been conducted in structured human environments, and as such is poorly adapted to unstructured forests. In this paper, we compare the performance of state of the art monocular SLAM systems on forest data and use visual appearance statistics to characterise the differences between forests and other environments, including a photorealistic simulated forest. We find that SLAM systems struggle with all but the most straightforward forest terrain and identify key attributes (lighting changes and in-scene motion) which distinguish forest scenes from "classic" urban datasets. These differences offer an insight into what makes forests harder to map and open the way for targeted improvements. We also demonstrate that even simulations that look impressive to the human eye can fail to properly reflect the difficult attributes of the environment they simulate, and provide suggestions for more closely mimicking natural scenes.

show abstract

Mapping forests using an unmanned ground vehicle with 3D LiDAR and graph-SLAM

Cited by 190 publications

References 27 publications

Extraction of Sample Plot Parameters from 3D Point Cloud Reconstruction Based on Combined RTK and CCD Continuous Photography

Extraction of Sample Plot Parameters from 3D Point Cloud Reconstruction Based on Combined RTK and CCD Continuous Photography

Comparing Three Different Ground Based Laser Scanning Methods for Tree Stem Detection

Visual Appearance Analysis of Forest Scenes for Monocular SLAM

Contact Info

Product

Resources

About