Processing coastal imagery with Agisoft Metashape Professional Edition, version 1.6—Structure from motion workflow documentation

Over, Jin‐Si R.; Ritchie, Andrew C.; Kranenburg, Christine J.; Brown, Jennifer; Buscombe, Daniel; Noble, Tom; Sherwood, C. R.; Warrick, Jonathan A.; Wernette, Phillipe

doi:10.3133/ofr20211039

Cited by 60 publications

(86 citation statements)

References 10 publications

(6 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Image resolution and outlier filtering are key parameters that can be adjusted during the SfM processing. A strong understanding of photogrammetric analysis principles can provide key insights into how these parameters may be adjusted to yield more successful 3D reconstructions (Over et al, 2021; USGS, 2017), but empirical validation of these workflows in the context of forest inventories is generally lacking. Only one study to our knowledge has evaluated image resolution and point cloud filtering parameters in the context of ITD (Tinkham & Swayze, 2021).…”

Section: Introductionmentioning

confidence: 99%

Optimizing aerial imagery collection and processing parameters for drone‐based individual tree mapping in structurally complex conifer forests

2022

View full text Add to dashboard Cite

1. Recent advances in remotely piloted aerial systems ('drones') and imagery processing enable individual tree mapping in forests across broad areas with lowcost equipment and minimal ground-based data collection. One such method involves collecting many partially overlapping aerial photos, processing them using 'structure from motion' (SfM) photogrammetry to create a digital 3D representation and using the 3D model to detect individual trees. SfM-based forest mapping involves myriad decisions surrounding methods and parameters for imagery acquisition and processing, but it is unclear how these individual decisions or their combinations impact the quality of the resulting forest inventories.2. We collected and processed drone imagery of a moderate-density, structurally complex mixed-conifer stand. We tested 22 imagery collection methods (altering flight altitude, camera pitch and image overlap), 12 imagery processing parameterizations (image resolutions and depth map filtering intensities) and 286 tree detection methods (algorithms and their parameterizations) to create 7,568 tree maps. We compared these maps to a 3.23-ha ground reference map of 1,775 trees >5 m tall that we created using traditional field survey methods.3. The accuracy of individual tree detection (ITD) and the resulting tree maps was generally maximized by collecting imagery at high altitude (120 m) with at least 90% image-to-image overlap, photogrammetrically processing images into a canopy height model (CHM) with a twofold upscaling (coarsening) step and detecting trees from the CHM using a variable window filter after applying a moving window mean smooth to the CHM. Using this combination of methods, we mapped trees with an accuracy exceeding expectations for structurally complex forests (for canopy-dominant trees >10 m tall, sensitivity = 0.69 and precision = 0.90). Remotely measured tree heights corresponded to groundmeasured heights with R 2 = 0.95. Accuracy was higher for taller trees and lower for understorey trees and would likely be higher in less dense and less structurally complex stands.

show abstract

Section: Introductionmentioning

confidence: 99%

Optimizing aerial imagery collection and processing parameters for drone‐based individual tree mapping in structurally complex conifer forests

2022

View full text Add to dashboard Cite

show abstract

“…To quantify the effectiveness of noise reduction in the sparse cloud, the RU parameter was used. A value of RU around 15 is equivalent to a B/H ratio of 1/5.5 (corresponding to an angle between homologous rays of about 20 degrees); a few studies suggest values between 10 and 15, unless this results in a TP removal greater than 50% [13]; a higher removal percentage can lead to an inaccurate model. The graph in Figure 4 shows on the y-axis the number of selected TP as a percentage of the total, as the RU parameter (x-axis) changes before the application of the masks (red dots) and after (orange squares); the trend of both curves is exponential.…”

Section: Resultsmentioning

confidence: 99%

A method to improve the accuracy of sparse UAV point cloud applied to the Amphitheater of Pompeii

Barba

Benedetto

Fiani

et al. 2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

In the field of cultural heritage dissemination, the availability of an accurate 3D model is mandatory to better describe an historically relevant building and provide the user with as much information as possible. In our paper we propose a workflow to improve the accuracy of the sparse cloud (Tie Points) by means of automated mask generation on photograms acquired by UAV (Unmanned Aerial Vehicle). In the first step, the statistical distribution of reprojection errors for each image, before applying the masks, is analyzed. In the second step, masks are automatically generated for each image to exclude, in the following alignment process, parts of the image characterized by reprojection error values outside the chosen confidence interval. The BIC criterion was used to identify the probability distribution that best fits the data. The results are promising and highlight the need to implement specific processes for preparing the input data to improve the accuracy of the output 3D model. Such results are not easily achievable with the processes implemented by default in the most widely used commercial photogrammetric software.

show abstract

“…Data set A (Sherwood et al, 2021) consists of one, three-band orthomosaic image (Figure 4a, Table 1), at 5-cm and also downsampled to a resolution of 25-cm, for mapping beach substrates of Sandwich Town Neck Beach on Cape Cod, Massachusetts. The orthomosaics are created from photographs collected from a low-altitude Uncrewed Aircraft System (UAS) on 21 September 2016, using a structure-from-motion workflow similar to that described by Over et al (2021) for high-resolution elevation mapping of coasts from aerial imagery (Warrick et al, 2019). The 5-cm and 25-cm pixel imagery are divided into 1,024 × 1,024 pixel, 3-band (RGB) tiles for annotation, which results in 99 and six tiles for the respective resolutions.…”

Section: Sedimentary Mapping Of a Mixed-sand-gravel Beach From Visibl...mentioning

confidence: 99%

“…The orthomosaics are created from photographs collected from a low‐altitude Uncrewed Aircraft System (UAS) on 21 September 2016, using a structure‐from‐motion workflow similar to that described by Over et al. (2021) for high‐resolution elevation mapping of coasts from aerial imagery (Warrick et al., 2019). The 5‐cm and 25‐cm pixel imagery are divided into 1,024 × 1,024 pixel, 3‐band (RGB) tiles for annotation, which results in 99 and six tiles for the respective resolutions.…”

Section: Datasets and Case Studiesmentioning

confidence: 99%

“…In other situations, there is a choice over what grid size to use when making the imagery, such as when converting from ungridded to gridded data. The orthomosaics are created from color‐attributed 3D point clouds (Over et al., 2021), therefore we use data set A (Table 1) to discuss a workflow designed to experimentally determine optimal grid size and image size ahead of a large labeling task.…”

Section: Datasets and Case Studiesmentioning

confidence: 99%

See 1 more Smart Citation

Human‐in‐the‐Loop Segmentation of Earth Surface Imagery

Buscombe

Goldstein

Sherwood³

et al. 2022

Earth and Space Science

Self Cite

View full text Add to dashboard Cite

Segmentation, or the classification of pixels (grid cells) in imagery, is ubiquitously applied in the natural sciences. Manual methods are often prohibitively time-consuming, especially those images consisting of small objects and/or significant spatial heterogeneity of colors or textures. Labeling complicated regions of transition that in Earth surface imagery are represented by collections of mixed-pixels, -textures, and -spectral signatures, can be especially error-prone because it is difficult to reliably unmix, identify and delineate consistently. However, the success of supervised machine learning (ML) approaches is entirely dependent on good label data. We describe a fast, semi-automated, method for interactive segmentation of N-dimensional (x, y, N) images into two-dimensional (x, y) label images. It uses human-in-the-loop ML to achieve consensus between the labeler and a model in an iterative workflow. The technique is reproducible; the sequence of decisions made by human labeler and ML algorithms can be encoded to file, so the entire process can be played back and new outputs generated with alternative decisions and/or algorithms. We illustrate the scientific potential of segmentation of imagery of diverse settings and image types using six case studies from river, estuarine, and open coast environments. These photographic and non-photographic imagery consist of 1-and 3-bands on regular and irregular grids ranging from centimeters to tens of meters. We demonstrate high levels of agreement in label images generated by several labelers on the same imagery, and make suggestions to achieve consensus and measure uncertainty, ideal for widespread application in training supervised ML for image segmentation.Plain Language Summary Labeling pixels in scientific images by hand is time-consuming and error-prone, so we would like to train computers to do that for us. We can use automated techniques from Artificial Intelligence or AI, like one called Deep Learning, but it needs a lot of example images and corresponding labels that have been made by hand. So, we still need to label quite a lot of images at the pixel level-called image segmentation. We made a computer program called Doodler that speeds up the process; you label some pixels, and it labels the rest. It is the fastest method we know of for image segmentation because it is semi-automated. We also show that it produces accurate and precise labeling, as we demonstrated by having multiple people use this method to label the same images. Because it is so fast and accurate, it allows us to get enough data to train Deep Learning models to do segmentation on all the images we have, from the past and in the future. Doodler therefore enables geoscientists to use Artificial Intelligence to extract much more information from their imagery, in service of geoscience in general.

show abstract

Processing coastal imagery with Agisoft Metashape Professional Edition, version 1.6—Structure from motion workflow documentation

Cited by 60 publications

References 10 publications

Optimizing aerial imagery collection and processing parameters for drone‐based individual tree mapping in structurally complex conifer forests

Optimizing aerial imagery collection and processing parameters for drone‐based individual tree mapping in structurally complex conifer forests

A method to improve the accuracy of sparse UAV point cloud applied to the Amphitheater of Pompeii

Human‐in‐the‐Loop Segmentation of Earth Surface Imagery

Contact Info

Product

Resources

About