3D Reconstruction of Plant/Tree Canopy Using Monocular and Binocular Vision (original) (raw)
Related papers
Plant or Tree Reconstruction Based on Stereo Vision
2013 Kansas City, Missouri, July 21 - July 24, 2013, 2013
Three dimensional (3D) reconstruction of the plant or tree canopy is an important step in order to measure canopy geometry, such as, height, width, volume, and leaf cover. In this research, binocular stereo vision was used to recover the 3D information of the canopy. A revised camera calibration method was provided to calibrate the cameras in world coordinate system. Only two images were used to realize a dense reconstruction. These two images were firstly rectified to make sure the corresponding feature points in the left and right images were on the same horizontal line. An efficient large scale stereo matching (ELAS) algorithm was used to find the disparity map. The plant or tree canopy was finally reconstructed based on these calibrated camera matrices and the disparity map through a triangulation method. A plant (croton) with big leaves and a small citrus tree with small leaves were used to test this two-view dense reconstruction. It was easy to measure the geometry of the big leaf. Two big leaves from croton plant were used to measure the width and length of the leaves. The measurement from the reconstruction and manual measurement showed that this reconstruction was metric reconstruction. Another three reconstructions were completed based on a side view of the croton plant, a top view of the croton plant, and a side view of the citrus tree. All these gave good 3D visualization of the objects.
Three-dimensional dense reconstruction of plant or tree canopy based on stereo vision
2018
Three-dimensional (3D) reconstruction of the plant or tree canopy is an important step to measure canopy geometry, volume, and leaf cover density for applications in precision agriculture, robotic harvesting, or plant phenotype. In this research, binocular stereo vision was used to recover the 3D information of the canopy. A revised camera calibration method was provided to calibrate the cameras in the world coordinate system. Only two images were used to realize a dense reconstruction. These two images were firstly rectified to make sure the corresponding feature points in the left and right images were on the same horizontal line. An efficient large-scale stereo matching (ELAS) algorithm was used to find the disparity map. The plant or tree canopy was finally reconstructed based on these calibrated camera matrices and the disparity map through a triangulation method. In this research, a series of laboratory experiments were conducted to validate the 3D reconstruction and verify th...
Assessment of tree structure using a 3D image analysis technique - A proof of concept
Efforts to improve the efficiency and efficacy of tree structure and crown architecture measurement are necessary to reduce error associated with indirect estimation of volume, which affects biophysical and ecosystem modelling, as well as resource assessment. In this short communication, we test the potential for a commercial SfM-MVS (Structure from Motion coupled with Multiple-View Stereophotogrammetry) software package, Photoscan-Professional, to accurately determine tree height, stem diameter, and eventually volume. SfM is a technique in computer vision, which calculates the 3D position of objects in a scene from a series of photographs. It uses a technique assuming that an object in a 3D scene is located on a vector between the image of the object in the camera and the object itself. The technique allows the construction of a 3D pointcloud, such as the one produced by laser technologies -terrestrial or airborne LiDAR. SfM requires no camera calibration or control points to construct an initial model. Moreover, it is a low-cost alternative to laser technologies. The second part of our method -MVS -is a visualization technique that reconstructs a 3D textured mesh of the scene from the SfM-derived pointcloud and RGB photographs. As a proof of concept, our methods were limited to two scenarios: (1) a single potted tree in a lab environment, where exact measurements could be made, and (2) two trees of different species and size in natural environments to test feasibility outside the laboratory. Precise measurements of tree height and stem diameter were compared with estimates obtained from the 3D model created using SfM-MVS. The results indicate that the SfM method is a promising -and inexpensive -alternative to terrestrial LiDAR and 3D scanners. Tree height estimates had error of 2.59%, while stem diameter estimates had error of 3.7%. The MVS algorithm used in this study was developed for plan surfaces such as topography or 'compact' objects and does not provide a representative 3D mesh for slender trees, although it works well for large stems. The authors link this disparity to the complex branching structure of trees.
A photogrammetry-based methodology to obtain accurate digital ground-truth of leafless fruit trees
Computers and Electronics in Agriculture
In recent decades, a considerable number of sensors have been developed to obtain 3D point clouds that have great potential in optimizing management in agriculture through the application of precision agriculture techniques. In order to use the data provided by these sensors, it is essential to know their measurement error. In this paper, a methodology is presented for obtaining a 3D point cloud of a central axis training system defoliated fruit tree (Malus domestica Bork.) obtained from stereophotogrammetry techniques based on structure-from-motion (SfM) and multi-view stereo-photogrammetry (MVS). The point cloud was made from a set of 288 photographs of the scene including the ground truth tree which was used to generate the digital 3D model. The resulting point cloud was validated and proven to faithfully represent reality. The bias of the resulting model is − 0.15 mm and 0.05 mm, for diameters and lengths, respectively. In addition, the presented methodology allows small changes in the ground truth actual tree to be detected as a consequence of the wood dehydration process. Having an actual and a digital ground-truth is the basis for validating other sensing systems for 3D vegetation characterization which can be used to obtain data to make more informed management decisions.
A digital photographic method for 3D reconstruction of standing tree shape
Annals of Forest Science, 2007
A digital photographic method is presented which is able to reconstruct the profile of the stem on standing trees up to a height of 12 m and to provide a fine level of detail. The method uses two digital photographs taken at 90 • to each other and does not require special illumination conditions. A method is proposed to perform the data acquisition process from the two photographs and to transform the stem dimensions and 3-D position from pixels to units of length. The accuracy of this method for measuring tree shape was tested by comparing the results with those obtained from a laser system. The comparison showed that the photographic method provides a good assessment of standing tree shape. 3D profile / standing trees / digital camera / photogrammetry / Picea abies Résumé-Mesure de la forme des arbres sur pied par photogrammétrie. La méthode photographique présentée dans cet article permet la reconstruction en 3D du profil du tronc d'arbres sur pied jusqu'à une hauteur de 12 m. On utilise un appareil photo numérique pour faire deux prises de vue orientées à 90 • l'une de l'autre. Il n'y a pas de contrainte particulière d'éclairage. Le traitement des images conduit à la représentation tridimensionnelle des tiges et à la mesure des indicateurs de forme. La précision de la méthode proposée est évaluée par référence à des mesures réalisées avec un théodolite à visée laser. Les résultats montrent que la méthode photographique permet des mesures précises de la forme des arbres. profil 3D / arbre sur pied / photo numérique / photogrammétrie / Picea abies
Fruit Trees 3D Data Acquisition and Reconstruction Based on Multi-source
Computer and Computing Technologies in Agriculture XI, 2019
In order to realize three-dimensional reconstruction of canopy at different growth stages of fruit trees, 3D data acquisition methods and canopy reconstruction methods were studied. Based on the analysis of morphological and structural changes in fruit phonological phase, and integrating the advantages of different data acquisition techniques, the data acquisition method of fruit tree morphological structure based on multi-source is proposed. In the dormant period, the canopy skeleton is extracted based on point cloud data; in the leaf curtain stage, a new artificial coding method of canopy structure is constructed, and the data of new shoots and leafs is obtained efficiently; and organ template data is obtained synchronously, and the organ template library is constructed. Then, a multi-source data fusion modeling method is proposed to reconstruct the three-dimensional canopy of fruit trees at different growth stages. And the feasibility of the method is verified by 12 year old open central leader system apple trees, the results show that compared with the manual data acquisition method, the method improves the efficiency by more than 5 times, and the error rate is less than 6%. It provides a feasible scheme for the continuous data acquisition and canopy 3D reconstruction of fruit trees, so as to provide technical support for virtual modeling, scientific calculation and experimental simulations.
Vertical Optical Scanning with Panoramic Vision for Tree Trunk Reconstruction
Sensors (Basel, Switzerland), 2017
This paper presents a practical application of a technique that uses a vertical optical flow with a fisheye camera to generate dense point clouds from a single planimetric station. Accurate data can be extracted to enable the measurement of tree trunks or branches. The images that are collected with this technique can be oriented in photogrammetric software (using fisheye models) and used to generate dense point clouds, provided that some constraints on the camera positions are adopted. A set of images was captured in a forest plot in the experiments. Weighted geometric constraints were imposed in the photogrammetric software to calculate the image orientation, perform dense image matching, and accurately generate a 3D point cloud. The tree trunks in the scenes were reconstructed and mapped in a local reference system. The accuracy assessment was based on differences between measured and estimated trunk diameters at different heights. Trunk sections from an image-based point cloud w...
A stereo imaging system for measuring structural parameters of plant canopies
Plant Cell and Environment, 2007
Plants constantly adapt their leaf orientation in response to fluctuations in the environment, to maintain radiation use efficiency in the face of varying intensity and incidence direction of sunlight. Various methods exist for measuring structural canopy parameters such as leaf angle distribution. However, direct methods tend to be labour-intensive, while indirect methods usually give statistical information on stand level rather than on individual leaves. We present an area-based, binocular stereo system composed of commercially available components that allows three-dimensional reconstruction of small- to medium-sized canopies on the level of single leaves under field conditions. Spatial orientation of single leaves is computed with automated processes using modern, well-established stereo matching and segmentation techniques, which were adapted for the properties of plant canopies, providing high spatial and temporal resolution (angle measurements with an accuracy of approx. ±5° and a maximum sampling rate of three frames per second). The applicability of our approach is demonstrated in three case studies: (1) the dihedral leaflet angle of an individual soybean was tracked to monitor nocturnal and daytime leaf movement showing different frequencies and amplitudes; (2) drought stress was diagnosed in soybean by quantifying changes in the zenith leaflet angle distribution; and (3) the diurnal course of the zenith leaf angle distribution of a closed soybean canopy was measured.
Proceedings of Computer Vision Problems in Plant Phenotyping, 2014
Functional-structural modeling and high-throughput phenomics demand tools for 3D measurements of plants. In this work, structure from motion is employed to estimate the position of a hand-held camera, moving around plants, and to recover a sparse 3D point cloud sampling the plants’ surfaces. Multiple-view stereo is employed to extend the sparse model to a dense 3D point cloud. The model is automatically segmented by spectral clustering, properly separating the plant’s leaves whose surfaces are estimated by fitting trimmed B-splines to their 3D points. These models are accurate snapshots for the aerial part of the plants at the image acquisition moment and allow the measurement of different features of the specimen phenotype. Such state-of-the-art computer vision techniques are able to produce accurate 3D models for plants using data from a single free moving camera, properly handling occlusions and diversity in size and structure for specimens presenting sparse canopies. A data set formed by the input images and the resulting camera poses and 3D points clouds is available, including data for sunflower and soybean specimens.
Camera Calibration for 3D Leaf-Image Reconstruction using Singular Value Decomposition
International Journal of Advanced Computer Science and Applications, 2017
Features of leaves can be more precisely captured using 3D imaging. A 3D leaf image is reconstructed using two 2D images taken using stereo cameras. Reconstructing 3D from 2D images is not straightforward. One of the important steps to improve accuracy is to perform camera calibration correctly. By calibrating camera precisely, it is possible to project measurement of distances in real world to the image plane. To maintain the accuracy of the reconstruction, the camera must also use correct parameter settings. This paper aims at designing a method to calibrate a camera to obtain its parameters and then using the method in the reconstruction of 3D images. Camera calibration is performed using region-based correlation methods. There are several steps necessary to follow. First, the world coordinate and the 2D image coordinate are measured. Extraction of intrinsic and extrinsic camera parameters are then performed using singular value decomposition. Using the available disparity image and the parameters obtained through camera calibration, 3D leafimage reconstruction can finally be performed. Furthermore, the results of the experimental depth-map reconstruction using the intrinsic parameters of the camera show a rough surface, so that a smoothing process is necessary to improve the depth map.