Terrestrial 3D laser scanning to track the increase in canopy height of both monocot and dicot crop species under field conditions
- Michael Friedli†1Email author,
- Norbert Kirchgessner†1,
- Christoph Grieder1,
- Frank Liebisch1,
- Michael Mannale1 and
- Achim Walter1
© Friedli et al. 2016
Received: 18 August 2015
Accepted: 12 December 2015
Published: 29 January 2016
Plant growth is a good indicator of crop performance and can be measured by different methods and on different spatial and temporal scales. In this study, we measured the canopy height growth of maize (Zea mays), soybean (Glycine max) and wheat (Triticum aestivum) under field conditions by terrestrial laser scanning (TLS). We tested the hypotheses whether such measurements are capable to elucidate (1) differences in architecture that exist between genotypes; (2) genotypic differences between canopy height growth during the season and (3) short-term growth fluctuations (within 24 h), which could e.g. indicate responses to rapidly fluctuating environmental conditions. The canopies were scanned with a commercially available 3D laser scanner and canopy height growth over time was analyzed with a novel and simple approach using spherical targets with fixed positions during the whole season. This way, a high precision of the measurement was obtained allowing for comparison of canopy parameters (e.g. canopy height growth) at subsequent time points.
Three filtering approaches for canopy height calculation from TLS were evaluated and the most suitable approach was used for the subsequent analyses. For wheat, high coefficients of determination (R2) of the linear regression between manually measured and TLS-derived canopy height were achieved. The temporal resolution that can be achieved with our approach depends on the scanned crop. For maize, a temporal resolution of several hours can be achieved, whereas soybean is ideally scanned only once per day, after leaves have reached their most horizontal orientation. Additionally, we could show for maize that plant architectural traits are potentially detectable with our method.
The TLS approach presented here allows for measuring canopy height growth of different crops under field conditions with a high temporal resolution, depending on crop species. This method will enable advances in automated phenotyping for breeding and precision agriculture applications. In future studies, the TLS method can be readily applied to detect the effects of plant stresses such as drought, limited nutrient availability or compacted soil on different genotypes or on spatial variance in fields.
KeywordsLaser scanning Scan point cloud Canopy height growth Maize Soybean Wheat
Plant growth is a good indicator of crop performance and is measureable by different methods and on different spatial and temporal scales . Plant growth reveals detailed information about the state of a plant  and allows for the assessment of the tolerance of a plant to abiotic stress such as drought [3, 4], heat  or nutrient deficiency . Today’s technologies offer many different possibilities to measure plant growth automatically, non-invasively and non-destructively. Many of these technologies construct a 3D scan point cloud of plants or canopies. A very simple approach to measure plant growth by taking images with a commercial digital camera was used for example by Sritarapipat et al.  to observe plant height changes in a rice field. In more complex approaches 3D images of plants are reconstructed by using stereo cameras , by analysing multiple images taken from different viewing angles  or by taking depth images . In recent studies, unmanned aerial vehicles (UAVs) were used to generate 3D reconstructions of winter wheat from multiple images to estimate crop height  and to generate 3D digital surface models of barley from hyperspectral information . In another study, a laser scanner was mounted on a UAV to estimate crop height of maize .
A very interesting and precise technology is the so-called 3D digitizer which uses ultrasonic or electromagnetic devices (digitizing pens) to construct 3D images of plant parts or whole plants. Plant architecture of different crops was measured with 3D digitizers to calculate light models in plant canopies in rice  and cucumber . 3D digitizing is very labour and time intensive because the digitizing pen needs to be manually pointed to important landmarks on the plant (for example leaf and shoot tips) to map plant architecture in 3D. Therefore, this technology cannot be used as an automated, high throughput phenotyping system.
A sophisticated technology, that is becoming more and more important, is the active remote-sensing laser rangefinder which uses a laser beam to determine the distance to an object. Different principles for distance detection exist (for a review see  or ).
Terrestrial laser scanning (TLS) offers a unique opportunity to make non-invasive and non-destructive measurements of canopies to characterize plant growth and to analyze diverse architectural parameters. TLS measurements render point clouds that depict the surface of the visible canopy oriented towards the observing device. These point clouds can be further analyzed, which has been done already in the fields of (1) forest ecology; (2) precision agriculture; and (3) phenotyping.
So far, most TLS studies were conducted in forest ecology to measure tree height, volume, leaf area, biomass and other important plant parameters [17–19]. Hosoi and Omasa  for example investigated the seasonal change of broad-leaved woody canopy leaf area density profiles. The plant structure and chlorophyll content in broadleaf saplings was studied by Eitel et al. .
In precision agriculture, measurements of orchard volumes [22, 23] or leaf area in orchards  or viticulture  have been conducted. The geometric characterization of tree crops is important for a number of different aspects such as the application of pesticides or irrigation systems (see  for a review). The aspect of canopy characterization is also important in vineyards to improve pesticide application methods . In field crops, TLS was applied to discriminate maize plants from weeds and soil for a targeted application of herbicides . Sensing of the nitrogen status of wheat plants by TLS was used for improved application of nitrogen fertilizers . In another approach, Saeys et al.  used TLS to estimate crop density of wheat that could be used to automatically adjust the speed of a combine harvester for a constant intake of biomass.
Another important research field in which TLS is applied is plant phenotyping under lab or field conditions. Morphological plant parameters such as canopy height [29, 30] and leaf area [31, 32] have been investigated. Besides morphological parameters also structural (number of leaves, orientation of surfaces, topology) and functional information (photosynthesis, stomatal conductance etc.) has been studied . Biomass [29, 34, 35] is probably the second most important parameter next to height growth . In a newer approach, the detection of individual maize plants has been performed to improve plant growth models or crop management strategies . The relevance of TLS measurements for field research remained limited though, since TLS measurements were typically conducted on single plants in pots [37, 38] or on small plants like Arabidopsis thaliana  from which conclusions to crops cannot easily be drawn. Furthermore, these measurements were often carried out under controlled and relatively artificial environmental conditions such as in climate chambers  or greenhouses [32, 40]. If TLS measurements were conducted in the field, this was done on very small areas ( 1 m2) or with a low resolution .
Therefore, we tested the hypotheses, whether TLS field measurements are capable to elucidate (1) differences in architecture that exist between genotypes; (2) genotypic differences in canopy height growth during the season and (3) short-term growth fluctuations (within 24 h), which could indicate e.g. responses to rapidly fluctuating environmental conditions.
Correlation between manually and TLS-derived wheat canopy height
The investigated ROIs were analyzed by testing three filtering approaches with different percentiles. The first FA used all contained scan points of each ROI and calculated certain percentiles of the z coordinate. The second FA filtered the ROI patchwise by a weighted median of certain percentiles. Thereby, the full ROI was considered, even if extreme values were present as for example in very heterogeneous field situations with a few very high plants. See “Data processing and data analysis” Sect. for a detailed description of the calculation of FAs. In the FA, which included all points of a ROI in the calculation (FAALLPOINTS), R2 for P100 was lowest for all measurement dates (Fig. 4a). This percentile included all maximum points and as a consequence also potential outliers that contributed to the low R2. The values of R2 for the first two measurement dates gradually decreased from P99 to P90. At these two dates, the canopy was not yet closed, so that the laser beam reached lower parts of the canopy. Thus, the values of the lower percentiles did not depict the maximum but lower parts of the canopy.
In the FA using the weighted median of the maxima of each patch in a ROI (FAMEDIANMAX), R2 values of P100 for the three dates were also slightly lower than the ones of P99. R2 values for the last measurement date did not markedly vary between P99 and P90. The values of R2 for the first two measurement dates decreased from P99 to P90 and were always lower for the first date.
In the FA using the weighted median of P99 of each patch in a ROI (FAMEDIANP99) the highest values for R2 were obtained for the last measurement. However, the values were clearly lower for this FA compared to the other two FAs. The values for R2 of the first two dates were very similar from P100 to P96 but then diverged with decreasing percentiles below P95. The values of R2 for the first measurement were more or less uniform throughout the entire tested range of R2, whereas the values of R2 for the second measurement date decreased towards lower percentiles.
For the three evaluated FAs, P99 resulted in the highest values for R2 throughout the season. Therefore, P99 was considered as the TLS measure which approximated canopy height most realistically. In the next step, FAs were compared with each other on the basis of P99 values of TLS-derived canopy height that were plotted against manually measured reference heights (Fig. 4d–f). For FAALLPOINTS, R2 increased from 0.86 for the first date to 0.92 for the last date (Fig. 4d). For FAMEDIANMAX, R2 increased from 0.85 for the first date to 0.93 for the last date (Fig. 4e). For FAMEDIANP99, R2 was comparable for the first (R2 = 0.73) and the second date (R2 = 0.73) but was higher (R2 = 0.76) for the third date (Fig. 4f). Using the three dates combined in a regression, FAALLPOINTS (R2 = 0.99) and FAMEDIANP99 (R2 = 0.98) reflect higher coefficients of determination than FAMEDIANMAX (R2 = 0.95). Therefore, FAALLPOINTS was used in subsequent calculations.
Short-term canopy height growth
Plant architectural traits
Deviations of transformed positions of spherical targets
Overview of the deviations of transformed positions of spherical targets throughout the measurement period
Average deviations of transformed positions (in m ± standard deviation)
Number of transformed sphere positions
0.0084 ± 0.0039
0.0042 ± 0.0031
0.0021 ± 0.0006
So far, most TLS studies conducted very detailed  and mostly indoor [39, 44] measurements or they were carried out on large areas in the field  with often low spatial resolution. Time intervals between measurements were often in the range of weeks if ever several scans were made and until now, to our knowledge, no one examined the temporal resolution limits of TLS on canopy height growth of crops in the field. With our approach, we fill the gap between these two extremes. We obtained a better resolution as for example in  and therefore increase the applicability of TLS for breeding-related phenotyping and precision agriculture. For breeding, many different genotypes planted often on relatively small plots of a few square meters need to be characterized with respect to their performance and their reaction towards alterations of environmental parameters [2, 46]. Further, our approach can be used as ground truth calibration method for new measurement systems from e.g. UAVs developed for precision agriculture [10, 11].  obtained an R2 of only 0.7 and a constant underestimation of 0.19 m of the plant height from the UAV-based data compared to manually measured plant heights. In  the height measurements from UAV even indoors had a measurement error of above 3.5 cm. UAVs produce wind (downwash) by themselves that can move plant canopies when flying at low altitudes. Therefore, a certain distance of UAVs from the plant canopy is needed to exclude any influence on the plant canopy. Increasing the flight altitude, however, decreases the resolution of measurements and thus also the accuracy of plant canopy reconstructions. Therefore, our high precision method (R2 of 0.92 for FAALLPOINTS for the last date and R2 of 0.99 for FAALLPOINTS using the three dates combined) could be used to calibrate such systems. The fixed position of the spherical targets during the whole season on aluminium rods and solid ground screws is new in plant science under field conditions and at the same time a simple approach, which leads to a high precision of the measurement and therefore allows for comparison of canopy parameters (e.g. canopy height growth) at subsequent time points. In our approach, the spherical targets define a coordinate system that is fixed during the complete season and that can be used to transform scan point clouds from different measurement dates into one and the same reference coordinate system. The small deviations of transformed positions of spherical targets are a strong evidence for the high accuracy of our measurement setup (Table 1). These deviations include technical measurement limitations of the laser scanner and potential movement of the positions of spherical targets throughout the season and are at the same time giving a value for the best achievable accuracy of our TLS approach. In our approach only the spherical coordinates have to be known and no additional expensive device such as a tachymeter or a GPS, as e.g. used by , is needed to measure the exact position of targets at each measurement date. Such measurement devices have their inherent technical resolution limits and sources of error that may result in an accumulation of inaccuracies during the season.
By evaluating different FAs for calculating canopy height from TLS data, we could show that in the FAALLPOINTS, the 99th percentile is best suited for computing the wheat canopy height (Fig. 4). This FA has, compared to other studies [29, 45], the advantage that outliers are excluded from calculations and thus the risk of over- or underestimation of the canopy height can be reduced. Another benefit of this FA, compared with FAMEDIANP99 is that no potentially important points are a priori excluded from the calculation. The temporal dynamic of the canopy is also clearly visible in the progress of the value of R2: a lower canopy is more “susceptible” for underestimation of the canopy height depending on the chosen percentile.
In a recent study, Hammerle et al.  investigated the effect of reduced point density of TLS crop surface models of wheat and rye. They examined the effect of a stepwise point reduction on the calculated canopy height from the maximum points or from the 90th percentile (P90) compared with the original resolution and a low resolution scan. However, their low resolution scans and simulated reduced point clouds had only 30–50 points per m2. For our purposes, this would have been by far a too low resolution to detect genotypic differences. Moreover, we could show for wheat (Fig. 4) that—in younger growth stages—the real canopy height will be underestimated by using P90 and that using absolute maximum points for canopy height calculation involves the risk of including outliers.
The temporal resolution that can be achieved with our approach depends on the scanned crop. For maize that only shows slight leaf movements, a temporal resolution of several hours can be achieved by scanning e.g. three times per day. In contrast, soybean exhibits a strong diel leaf movement, resulting in a strong change of canopy height during a day (Fig. 8). Thus, soybean is ideally scanned only once per day, after leaves have reached their most horizontal orientation. By taking the difference of the canopy height between 2 days at this time point, the real increase in canopy height will be detected. Otherwise, detected changes in canopy height are a jumble of daily leaf movements and real canopy height growth. Maize and soybean illustrate how important a sound knowledge about physiological processes of a scanned crop is. Growth of maize (Fig. 6) and soybean (Fig. 7c, d) follows temperature. For maize this is no surprise, as it is known from literature, that growth of monocot species follows temperature (e.g. ). For soybean, the observed growth pattern with the highest growth in the afternoon is in contrast to the notion that dicot species show their maximal growth activity in the beginning of the day (type 1) or at the end of the day (type 2) . For soybean it was shown in several studies (e.g. [48, 49]) that maximal leaf growth occurs towards the end of the night.
For maize, we could show that plant architectural traits are detectable by TLS with our method (Fig. 9). The obtained scan point height distribution histograms indicate the height position of plant organs, such as leaves and ears and genotypic differences in light penetration properties as potentially affected by number of leaves, leaf area index or leaf angles. Neither for maize (Fig. 6) nor for soybean (Fig. 7) different growth patterns could be detected by TLS for different genotypes in this study. For the precise distinction of genotypes by TLS beyond canopy height detection further studies including more genotypic variance and more measurement points during the season will be needed.
With our TLS approach of data acquisition and data analysis, we established—compared to other TLS studies—a quite simple way of handling TLS data of field crops. Our TLS approach therefore can be considered as a valuable tool to measure the canopy height growth of different crops under field conditions. The high correlation between manually measured and TLS-derived canopy height of wheat is showing the high accuracy of our method (Fig. 4). The fact, that we could measure the diurnal pattern of canopy height growth in maize is another strong evidence for the accuracy of our TLS method (Fig. 6). However, there are some restrictions to perform meaningful measurements. No measurement can be conducted if it rains due to the laser scanner that is not completely weatherproof and also due to technical issues regarding the scattering of the laser beam on raindrops. However, many devices used for field phenotyping cannot be used during rain. During the scanning process, it should ideally not be windy to prevent a blurred point cloud of the scanned crop. The dependence on windless conditions, however, depends on the scanned crop and also on the developmental stage of the crop and the research question. The scanning of crops that are small, stiff and have less surface exposed to wind is less dependent on wind conditions. The scanning of younger and thus normally smaller plants is also less affected by wind. Wind speeds of 2 m s−1 are feasible as our approach uses the statistical percentile method and therefore has a certain robustness against deviations caused by wind.
The TLS approach presented here allows for measuring canopy height growth and architecture of different crops under field conditions with a high temporal resolution, depending on crop species. The approach will therefore be a valuable component of plant breeding programs. It can also facilitate the elucidation of stress-related plant responses in the field in a variety of plants. Furthermore, additional and new plant/crop parameters as for example canopy volume, leaf angle distribution (in the absence of wind) and height positions of key organs such as leaves and ears could be computed and analyzed by accordingly adjusting the scanning resolution and the distances between scanning positions.
Overview of the scanned species, dates, measurements per date, scan parameters and reference measurements
TLS measurements per date
Point distance at 2 m (in mm)
Point distance at 10 m (in mm)
23 Sep. 2013
27 Mar. 2014
02 Apr. 2014
09 Apr. 2014
15 Apr. 2014
09 May 2014
21 Jun. 2014
22 Jun. 2014
16 Jul. 2014
17 Jul. 2014
21 Jun. 2014
22 Jun. 2014
16 Jun. 2014
17 Jul. 2014
Setup in the field and data acquisition
Measurements were conducted in the field of the research station for plant science of ETH Zurich in Eschikon, Lindau in 2013 and 2014 (Table 2). Maize (Zea mays) and wheat (Triticium aestivum) as monocot species as well as soybean (Glycine max) as a dicot species were scanned periodically with the laser scanner throughout the season.
Fields were scanned from different positions at the same point in time in regular intervals ranging from several scans per day to weekly scans. White spherical targets (For maize and soybean: 14.5 cm in diameter, Laserscanning Europe GmbH, 39,120 Magdeburg, Germany; For wheat: 30 cm in diameter; do-it-yourself product) were distributed in the scanned area to allow for the later merging of the single scans from the same point in time but from different positions of a field to a scan point cloud. These targets were mounted on aluminium rods (Fig. 1c; 1.52 m in length for soybean and wheat; 3.02 m in length for maize; 3 cm in diameter for all crops) that in turn were fixed to ground screws (80 cm in length, Krinner GmbH, 3272 Walperswil, Switzerland). As spherical targets and aluminium rods are sensitive to environmental influences, they were only positioned in the field during measurement times. The ground screws were positioned within the rows to avoid any contact to machines. Thus, the position of the spherical targets remained constant during the season and defined a fixed coordinate system for all measurements. By transforming scan point clouds to this fixed coordinate system, scan point clouds throughout the season could be aligned for each crop, facilitating the comparison of the canopy at the different measurement points.
Overview of the cut parts and cutting points in the “height level experiment”
Tassel and flag leaf
Above second leaf
Above third leaf
Above forth leaf
Above the ear
Above the ear
Below the ear
In 2014, the scanned part of the maize field had a size of 6 m by 6 m and consisted of eight rows with a row spacing of 0.75 m. The two varieties ‘Bonfire’ and ‘Gottardo’ (KWS Saat SE, Einbeck, Germany) were sown each in four rows (Fig. 2b). Five white spherical targets were distributed over the scanned area and at each date, scans from the four corners were carried out at around 6 a.m., 1 p.m. and 7 p.m. (6 p.m. in July). The scanned part of the soybean field had a size of 6 m by 6 m and consisted of four plots with a size of 1.5 m by 6 m. The two varieties ‘Gallec’ (DSP, Delley Switzerland) and ‘Lissabon’ (fenaco Genossenschaft, Bern, Switzerland) were sown each on two plots (Fig. 2c). Five white spherical targets were distributed over the scanned area and at each date, scans from the four corners were conducted at around 8 am, 3 p.m. and 9 p.m. (8 p.m. in July). The wheat field had a size of around 30 m by 40 m (Fig. 2a). Seven white spherical targets were distributed over the whole wheat field and the field was scanned from 16 positions distributed homogeneously over the field. For later analysis only a part of the field (around 24 m by 24 m), including 192 plots with each a size of 1.5 m by 1.7 m, was used. 156 different genotypes were sown in these 192 plots (see  for more details).
Manual height reference measurements in soybean and maize in 2014 were taken during the first and the last scan on ten plants per genotype. In maize, the distance from a nail head in the soil next to the plant and the tip of the youngest leaf, which was manually straightened into an upright position, was measured. In soybean the distance from a nail head in the soil next to the plant and the tip of the shoot axis was measured. Manual height reference measurements per plot in wheat in 2014 were taken by holding a yardstick in the canopy at three positions and reading the value.
Data processing and data analysis
- 1.Point cloud transformation to the fixed coordinate system: Sphere coordinates of all scans were manually exported to txt files. They were used to estimate the rigid coordinate transformation for all point clouds to fixed coordinates . The deviations of transformed sphere coordinates of each sphere from different scans were saved to txt-files as they give a value for the best achievable accuracy. The scan point clouds were then transformed to the fixed coordinate system.
(in case of soil level measurement) Generation of soil height as a “height image” (HS) with 5 mm pixel size (Fig. 3). Therefore the height minimum was first determined on the pixel grid. Gaps were interpolated. The result was median-filtered with a patch size of 21 cm.
(in case of plant height measurement) Subtraction of soil level from each point of the point cloud using the appropriate entry of HS. The appropriate pixels of HS were found by projection of the point cloud along the z-axis. All points which are projected on the same pixel belong to a column with quadratic base area of 5 mm × 5 mm. They were processed together in the further evaluation by calculating their percentiles (Fig. 3).
Selection of regions of interest (ROIs) as individual areas or as a grid (Fig. 2). Border rows were not within the selected ROIs to exclude border effects.
- 3.Height analyses of the point cloud were carried out with two classes of analysis approaches consisting of three filtering approaches (FAs) in total. The first class consisted of one FA (FAALLPOINTS) and the second class of two FAs (FAMEDIANMAX and FAMEDIANP99), respectively:
- (a)Percentiles of all points (FAALLPOINTS)
Calculation of the percentiles for each ROI by taking every point of the whole scan point cloud within the ROI into account.
Taking the maxima [or the xth percentile (Px)] for each 5 mm × 5 mm pixel, a height map HP of the point cloud was calculated (Fig. 3). Points higher than 10 cm were regarded as “plant points”, lower points were neglected.
Patchwise calculation of percentiles of HP (edge length 15 cm in direction of the row and full row width for maize and 10 cm × 10 cm for soybean and wheat). Number of plant pixels N per patch was saved for later “weighting”.
Calculation of the weighted median of percentiles per ROI. To calculate the weighted median of a percentile of a ROI each Px of all patches of the ROI was put N times to a list of which the median was calculated. Therefore patches with high plant coverage were stronger weighted than those with low plant coverage. In this study we applied the maximum value and P99 filter (FAMEDIANMAX and FAMEDIANP99).
Calculation of canopy height and growth or other parameters
MF, CG, FL, MM and AW designed the experiments. Experiments were performed by MF, CG and MM. NK designed the measurement concept and implemented the analysis process in MATLAB®. MF drafted the manuscript with help and contributions by NK, FL and AW. Analyses, figures and tables were mainly performed by MF and partially by MM and NK. All authors read and approved the final manuscript.
The authors would like to thank all our colleagues who supported us during the design of the software, data acquisition and analyses. Thank you also to Hansueli Zellweger for seeding, raising and taking care of the crops at the ETH research station for plant sciences in Eschikon, Lindau-. MF acknowledges support from the Swiss National Science Foundation (Grant nr. 315230_144078/1). CG acknowledges support from the European Union’s Seventh Framework Program for research, technological development and demonstration under grant agreement no GA-2010 267243 PLANT FELLOWS.
The authors declare that they have no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Walter A, Silk WK, Schurr U. Environmental effects on spatial and temporal patterns of leaf and root growth. Annu Rev Plant Biol. 2009;60:279–304.PubMedView ArticleGoogle Scholar
- Walter A, Liebisch F, Hund A. Plant phenotyping: from bean weighing to image analysis. Plant Methods. 2015;11:14.PubMedPubMed CentralView ArticleGoogle Scholar
- Banziger M, Edmeades GO, Lafitte HR. Selection for drought tolerance increases maize yields across a range of nitrogen levels. Crop Sci. 1999;39:1035–40.View ArticleGoogle Scholar
- Lipiec J, Doussan C, Nosalewicz A, Kondracka K. Effect of drought and heat stresses on plant growth and yield: a review. Int Agrophys. 2013;27:463–77.View ArticleGoogle Scholar
- Ha S, Tran LS. Understanding plant responses to phosphorus starvation for improvement of plant tolerance to phosphorus deficiency by biotechnological approaches. Crit Rev Biotechnol. 2014;34:16–30.PubMedView ArticleGoogle Scholar
- Sritarapipat T, Rakwatin P, Kasetkasem T. Automatic rice crop height measurement using a field server and digital image processing. Sensors. 2014;14:900–26.PubMedPubMed CentralView ArticleGoogle Scholar
- Biskup B, Scharr H, Schurr U, Rascher U. A stereo imaging system for measuring structural parameters of plant canopies. Plant Cell Environ. 2007;30:1299–308.PubMedView ArticleGoogle Scholar
- Paproki A, Sirault X, Berry S, Furbank R, Fripp J. A novel mesh processing based technique for 3D plant analysis. BMC Plant Biol. 2012;12:63.PubMedPubMed CentralView ArticleGoogle Scholar
- Chéné Y, Belin E, Rousseau D, Chapeau-Blondeau F. Multiscale analysis of depth images from natural scenes: scaling in the depth of the woods. Chaos Soliton Fract. 2013;54:135–49.View ArticleGoogle Scholar
- Khanna R, Möller M, Pfeifer J, Liebisch F, Walter A, Siegwart R. Beyond point clouds—3D mapping and field parameter measurements using UAVs. International Conference on Emerging Technologies and Factory Automation. vol 20. Luxembourg: IEEE; 2015.Google Scholar
- Aasen H, Burkart A, Bolten A, Bareth G. Generating 3D hyperspectral information with lightweight UAV snapshot cameras for vegetation monitoring: from camera calibration to quality assurance. ISPRS J Photogramm. 2015;108:245–59.View ArticleGoogle Scholar
- Anthony D, Elbaum S, Lorenz A, Detweiler C. On crop height estimation with UAVs. IEEE Int C Int Robot. 2014:4805–12.Google Scholar
- Zheng BY, Shi LJ, Ma YT, Deng QY, Li BG, Guo Y. Comparison of architecture among different cultivars of hybrid rice using a spatial light model based on 3D digitising. Funct Plant Biol. 2008;35:900–10.View ArticleGoogle Scholar
- Wiechers D, Kahlen K, Stutzel H. Evaluation of a radiosity based light model for greenhouse cucumber canopies. Agr For Meteorol. 2011;151:906–15.View ArticleGoogle Scholar
- Rosell JR, Sanz R. A review of methods and applications of the geometric characterization of tree crops in agricultural activities. Comput Electron Agr. 2012;81:124–41.View ArticleGoogle Scholar
- Omasa K, Hosoi F, Konishi A. 3D lidar imaging for detecting and understanding plant responses and canopy structure. J Exp Bot. 2007;58:881–98.PubMedView ArticleGoogle Scholar
- Parker GG, Harding DJ, Berger ML. A portable LIDAR system for rapid determination of forest canopy structure. J Appl Ecol. 2004;41:755–67.View ArticleGoogle Scholar
- Hopkinson C, Chasmer L, Young-Pow C, Treitz P. Assessing forest metrics with a ground-based scanning lidar. Can J Forest Res. 2004;34:573–83.View ArticleGoogle Scholar
- Clawges R, Vierling L, Calhoon M, Toomey M. Use of a ground-based scanning lidar for estimation of biophysical properties of western larch (Larix occidentalis). Int J Remote Sens. 2007;28:4331–44.View ArticleGoogle Scholar
- Hosoi F, Omasa K. Detecting seasonal change of broad-leaved woody canopy leaf area density profile using 3D portable LIDAR imaging. Funct Plant Biol. 2009;36:998–1005.View ArticleGoogle Scholar
- Eitel JUH, Vierling LA, Long DS. Simultaneous measurements of plant structure and chlorophyll content in broadleaf saplings with a terrestrial laser scanner. Remote Sens Environ. 2010;114:2229–37.View ArticleGoogle Scholar
- Rosell JR, Llorens J, Sanz R, Arno J, Ribes-Dasi M, Masip J, et al. Obtaining the three-dimensional structure of tree orchards from remote 2D terrestrial LIDAR scanning. Agr Forest Meteorol. 2009;149:1505–15.View ArticleGoogle Scholar
- Polo JRR, Sanz R, Llorens J, Arno J, Escola A, Ribes-Dasi P, et al. A tractor-mounted scanning LIDAR for the non-destructive measurement of vegetative volume and surface area of tree-row plantations: a comparison with conventional destructive measurements. Biosyst Eng. 2009;102:128–34.View ArticleGoogle Scholar
- Arno J, Escola A, Valles JM, Llorens J, Sanz R, Masip J, et al. Leaf area index estimation in vineyards using a ground-based LiDAR scanner. Precis Agric. 2013;14:290–306.View ArticleGoogle Scholar
- Llorens J, Gil E, Llop J, Escola A. Ultrasonic and LIDAR sensors for electronic canopy characterization in vineyards: advances to improve pesticide application methods. Sensors. 2011;11:2177–94.PubMedPubMed CentralView ArticleGoogle Scholar
- Andujar D, Rueda-Ayala V, Moreno H, Rosell-Polo JR, Escola A, Valero C, et al. Discriminating crop, weeds and soil surface with a terrestrial LIDAR sensor. Sensors. 2013;13:14662–75.PubMedPubMed CentralView ArticleGoogle Scholar
- Eitel JUH, Vierling LA, Long DS, Hunt ER. Early season remote sensing of wheat nitrogen status using a green scanning laser. Agr Forest Meteorol. 2011;151:1338–45.View ArticleGoogle Scholar
- Saeys W, Lenaerts B, Craessaerts G, De Baerdemaeker J. Estimation of the crop density of small grains using LiDAR sensors. Biosyst Eng. 2009;102:22–30.View ArticleGoogle Scholar
- Tilly N, Hoffmeister D, Cao Q, Huang SY, Lenz-Wiedemann V, Miao YX, et al. Multitemporal crop surface models: accurate plant height measurement and biomass estimation with terrestrial laser scanning in paddy rice. J Appl Remote Sens. 2014;8(1):083671.View ArticleGoogle Scholar
- Lumme J, Karjalainen M, Kaartinen H, Kukko A, Hyyppä J, Hyyppä H et al. Terrestrial laser scanning of agricultural crops. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. 2008;XXXVII Part B5:563–6.Google Scholar
- Gebbers R, Ehlert D, Adamek R. Rapid mapping of the leaf area index in agricultural crops. Agron J. 2011;103:1532–41.View ArticleGoogle Scholar
- Hosoi F, Nakabayashi K, Omasa K. 3-D modeling of tomato canopies using a high-resolution portable scanning lidar for extracting structural information. Sensors. 2011;11:2166–74.PubMedPubMed CentralView ArticleGoogle Scholar
- Sirault X, Fripp J, Paproki A, Kuffner P, Nguyen C, Li R et al. PlantScan™: a three-dimensional phenotyping platform for capturing the structural dynamic of plant development and growth. Proceedings of the 7th international conference on functional-structural plant models. 2013:45–8.Google Scholar
- Ehlert D, Adamek R, Horn HJ. Laser rangefinder-based measuring of crop biomass under field conditions. Precis Agric. 2009;10:395–408.View ArticleGoogle Scholar
- Eitel JUH, Magney TS, Vierling LA, Brown TT, Huggins DR. LiDAR based biomass and crop nitrogen estimates for rapid, non-destructive assessment of wheat nitrogen status. Field Crop Res. 2014;159:21–32.View ArticleGoogle Scholar
- Hofle B. Radiometric correction of terrestrial LiDAR point cloud data for individual maize plant detection. IEEE Geosci Remote Sens Lett. 2014;11:94–8.View ArticleGoogle Scholar
- Dornbusch T, Wernecke P, Diepenbrock W. A method to extract morphological traits of plant organs from 3D point clouds as a database for an architectural plant model. Ecol Model. 2007;200:119–29.View ArticleGoogle Scholar
- Paulus S, Dupuis J, Riedel S, Kuhlmann H. Automated analysis of barley organs using 3D laser scanning: an approach for high throughput phenotyping. Sensors. 2014;14:12670–86.PubMedPubMed CentralView ArticleGoogle Scholar
- Kaminuma E, Heida N, Tsumoto Y, Yamamoto N, Goto N, Okamoto N, et al. Automatic quantification of morphological traits via three-dimensional measurement of Arabidopsis. Plant J. 2004;38:358–65.PubMedView ArticleGoogle Scholar
- Kjaer KH, Ottosen CO. 3D laser triangulation for plant phenotyping in challenging environments. Sensors. 2015;15:13533–47.PubMedPubMed CentralView ArticleGoogle Scholar
- Hosoi F, Omasa K. Estimating vertical plant area density profile and growth parameters of a wheat canopy at different growth stages using three-dimensional portable lidar imaging. ISPRS J Photogramm. 2009;64:151–8.View ArticleGoogle Scholar
- Hoffmeister D, Waldhoff G, Curdt C, Tilly N, Bendig J, Bareth G. Spatial variability detection of crop height in a single field by terrestrial laser scanning. Precision Agriculture ‘13. Wageningen: Wageningen Academic Publishers; 2013. p. 267–74.Google Scholar
- Paulus S, Dupuis J, Mahlein AK, Kuhlmann H. Surface feature based classification of plant organs from 3D laserscanned point clouds for plant phenotyping. BMC Bioinformatics. 2013;14:238.PubMedPubMed CentralView ArticleGoogle Scholar
- Paulus S, Schumann H, Kuhlmann H, Leon J. High-precision laser scanning system for capturing 3D plant architecture and analysing growth of cereal plants. Biosyst Eng. 2014;121:1–11.View ArticleGoogle Scholar
- Hammerle M, Hofle B. Effects of reduced terrestrial LiDAR point density on high-resolution grain crop surface models in precision agriculture. Sensors. 2014;14:24212–30.PubMedPubMed CentralView ArticleGoogle Scholar
- White JW, Andrade-Sanchez P, Gore MA, Bronson KF, Coffelt TA, Conley MM, et al. Field-based phenomics for plant genetics research. Field Crop Res. 2012;133:101–12.View ArticleGoogle Scholar
- Poiré R, Wiese-Klinkenberg A, Parent B, Mielewczik M, Schurr U, Tardieu F, et al. Diel time-courses of leaf growth in monocot and dicot species: endogenous rhythms and temperature effects. J Exp Bot. 2010;61:1751–9.PubMedPubMed CentralView ArticleGoogle Scholar
- Ainsworth EA, Walter A, Schurr U. Glycine max leaflets lack a base-tip gradient in growth rate. J Plant Res. 2005;118:343–6.PubMedView ArticleGoogle Scholar
- Friedli M, Walter A. Diel growth patterns of young soybean (Glycine max) leaflets are synchronous throughout different positions on a plant. Plant, Cell Environ. 2015;38:514–24.View ArticleGoogle Scholar
- Faro. Phase shift measurement and time of flight measurement. 2011. http://www.faro.com/en-sg/resource-centre-ap/white-papers. Accessed 10 July 2015.
- Kirchgessner N, Liebisch F, Hund A, Walter A. Field imaging platform (FIP)—an automated system for plant phenotyping in the field. 20. und 21. Workshop Computer-Bildanalyse in der Landwirtschaft - 3. Workshop Unbemannte autonom fliegende Systeme (UAS) in der Landwirtschaft, Bornimer Agrartechnische Berichte 88:74.Google Scholar
- Grieder C, Hund A, Walter A. Image based phenotyping during winter: a powerful tool to assess wheat genetic variation in growth response to temperature. Funct Plant Biol. 2015;42:387–96.View ArticleGoogle Scholar
- Taati B. EstimateRigidTransform. MATLAB Central File Exchange. 2010. http://www.mathworks.com/matlabcentral/fileexchange/28305-estimaterigidtransform. Accessed 22 July 2014.