Smathermather's Weblog

Remote Sensing, GIS, Ecology, and Oddball Techniques

Archive for the ‘3D’ Category

OpenDroneMap on the road part II

Posted by smathermather on March 28, 2017

Thinking a little more about moderately large compute resources and their container (see previous post), I revised my analysis to look and see if we can fit these 10 NUCs plus switch and outlets into a carry-on sized case. It turns out, at first blush, it seems feasible:


Posted in 3D, Docker, OpenDroneMap, PDAL | Tagged: , , | 5 Comments »

OpenDroneMap on the road

Posted by smathermather on March 27, 2017


This is a theoretical post. Imagine for a moment that OpenDroneMap can scale to the compute resources that you have in an elastic and sane way (we are short weeks away from the first work on this), and so, if you are a typical person in the high-speed internet world, you might be thinking, “Great! Let’s throw this up on the cloud!”

But imagine for a moment you are in a network limited environment. Do you process on a local laptop? Do you port around a desktop? The folks in the humanitarian space think about this a lot — depending on the project, one could spend weeks or months in network limited environments.

Enter POSM

Folks at American Red Cross (ARC) have been thinking about this a lot. What has resulted, in order to aid in mapping e.g. rural areas in West Africa is Portable OpenStreetMap, or POSM, a tool for doing all the OpenStreetMap stuff, but totally and temporarily offline.

The software for this is critical, but I’ve been increasingly interested in the hardware side of things. OpenDroneMap, even with it’s upcoming processing, memory, and scaling improvements will still require more compute resources than, say OpenMapKit and Field Papers. I’ve been contemplating that once the improvements are in place, what kind of compute center could you haul in the field with you?

I’m not just thinking humanitarian and development use cases either — what can we do to make processing drone imagery in the field faster? Can we make it fast enough to get results before leaving the field? Can we modify our flight planning based on the stream of data being processed and adapt while we are there? Our real costs for flying are often finding staff and weather windows that are good, and sometimes we miss opportunities in the delay between imagery capture and processing. How can we close that loop faster?


On the hardware side of the house, the folks at ARC are using Intel NUC kits. For ODM, as I understand it, they go a step up in processing power from their specs to something with an i7. So, I got to thinking — can we put together a bunch of these, running on a generator, and not break the bank on weight (keep it under 50 lbs)? It turns out, maybe we can. For a round $10,000, you might assemble 10 of these 4-core NUCs with a network switch, stuff it into a Pelican Air 1605 case, with 320 GB RAM, and 2.5 TB of storage. More storage can be added if necessary.

This is a thought experiment so far, and may not be the best way to get compute resources in the field, your mileage may vary, etc., but it’s and interesting though.field_compute.PNG

Cost Breakdown


Follow up

Any thoughts? Anyone deployed serious compute resources to the field for drone image processing? I’d love to hear what you think.

Posted in 3D, Docker, OpenDroneMap, PDAL | Tagged: , , | 2 Comments »

Time for localization?

Posted by smathermather on March 26, 2017

Just saw this great blog post by my friend Mr. Yu at Korea National Park on using OpenDroneMap. If you need it in English, google seems to translate it rather well:

Maybe it’s time to look at localization for WebODM… .

Posted in 3D, OpenDroneMap, Other | Tagged: | Leave a Comment »

Scaling OpenDroneMap, necessary (and fun!) next steps

Posted by smathermather on March 8, 2017

Project State

OpenDroneMap has really evolved since I first put together a concept project presented at FOSS4G Portland in 2014, and hacked with my first users (Michele M. Tobias & Alex Mandel). At this stage, we have a really nicely functioning tool that can take drone images and output high-quality geographic products. The project has 45 contributors, hundreds of users, and a really great community (special shout-out to Piero Toffanin and Dakota Benjamin without whom the project would be nowhere near as viable, active, or wonderful). Recent improvements can be roughly categorized into data quality improvements and usability improvements. Data quality improvements were aided by the inclusion better point cloud creation from OpenSfM and better texturing from mvs-texturing. Usability improvements have largely been in the development of WebODM as a great to use and easy-to-deploy front end for OpenDroneMap.

With momentum behind these two directions — improved usability and improved data output, it’s time to think a little about how we scale OpenDroneMap. It works great for individual flights (up to a few hundred images at a time), but a promise of open source projects is scalability. Regularly we get questions from the community about how they can run ODM on larger and larger datasets in a sustainable and elastic way. To answer these questions, let me outline where we are going.

Project Future

Incremental optimizations

When I stated that scalability is one of the promises of open source software. I mostly meant scaling up: if I need more computing resources with an open source project, I don’t have to purchase more software licenses, I just need to rent or buy more computing resources. But an important element to scalability is the per unit use of computing resources as well. If we are not efficient and thoughtful about how we use things on the small scale, then we are not maximizing our scaled up resources.  Are we efficient in memory usage; is our matching algorithm as accurate as possible for the level of accuracy thus being efficient with the processor resources I have; etc.? I think of this as improving OpenDroneMap’s ability to efficiently digest data.

Magic school bus going doing the digestive system

Incremental toolchain optimizations are thus part of this near future for OpenDroneMap (and by consequence OpenSfM, the underlying computer vision tools for OpenDroneMap), focusing on memory and processor resources. The additional benefit here is that small projects and small computing resources also benefit. For humanitarian and development contexts where compute and network resources are limiting, these incremental improvements are critical. Projects like American Red Cross’ Portable OpenStreetMap (POSM) will benefit from these improvements, as will anyone in the humanitarian and development communities that need efficient processing of drone imagery offline.

To this end, three approaches are being considered for incremental improvements.  Matching speed could be improved by the use of Cascade Hashing matching or Bag of Words based method.Memory improvements could come via improved correspondence graph data structures and possibly SLAM-like pose-graph methods for global adjustment of camera positions in order to avoid global bundle adjustment.

Figure from Bag of Words paper

Figure from Bag of Words paper

Large-scale pipeline

In addition to incremental improvements, for massive datasets we need an approach to splitting up our dataset into manageable chunks. If incremental improvements help us better and more quickly process datasets, the large-scale pipeline is the teeth of this approach — we need to cut and chew up our large datasets into smaller chunks to digest.

Image of Dr. Teeth of the Muppets.

Dr. Teeth

If for a given node I can process 1000 images efficiently, but I have 80,000 images, I need a process that splits my dataset into 80 manageable chunks and processes through them sequentially or in parallel until done. Maybe I have 9000 images? Then I need it split into 9 chunks.

Image over island showing grid of 9 for spliting an aerial dataset

Eventually, I want to synthesize the outputs back into a single dataset. Ideally I split the dataset with some overlap as follows:

Image over island showing grid of 9 for spliting an aerial dataset shown with overlap

Problems with splitting SfM datasets

We do run into some very real problems with splitting our datasets into chunks for processing. There are a variety of issues, but the most stark is consistency issues from the resultant products. Quite often our X, Y, and Z values won’t match in the final reconstructions. This becomes critical when performing, e.g. hydrologic analyses on resultant Digital Terrain Models.

Water flow on patched DEM showing pooling effects around discontinuities

Water flow on patched DEM showing pooling effects around discontinuities (credit: Anna Petrasova et al)

Anna Petrasova et al address merging disparate DEM’s in GRASS with Seamless fusion of high-resolution DEMs from multiple sources with r.patch.smooth.

Water flow on fused DEM

Water flow on fused DEM showing corrected flow (credit: Anna Petrasova et al)

What Anna describes and solves is the problem of matching LiDAR and drone data and assumes that the problems between the datasets are sufficiently small that smoothing the transition between the datasets is adequate. Unfortunately, when we process drone imagery in chunks, we can get translation, rotation, skewing, and a range of other differences that often cannot be accounted for when we’re processing the digital terrain model at the end.

What follows is a small video of a dataset split and processed in two chunks. Notice offsets, rotations, and other issues of mismatch in the X and Y dimensions, and especially Z.

When we see these differences in the resultant digital terrain model, the problem can be quite stark:

Elevation differences along seamline of merged OpenDroneMap DTMs

Elevation differences along seamline of merged OpenDroneMap DTMs

To address these issues we require both the approach that Anna proposes that fixes for and smooths out small differences, and a deeper approach specific to matching drone imagery datasets to address the larger problems.

Deeper approach to processing our bites of drone data

To ensure we are getting the most out of stitching these pieces of data back together at the end, we require using a very similar matching approach to what we use in the matching of images to each other. Our steps will be something like as follows:

  • Split our images to groups
  • Run reconstruction on each group
  • Align and tranform those groups to each other using matching features between the groups
  • For secondary products, like Digital Terrain Models, blend the outputs using an approach similar to r.patch.smooth.

In close

I hope you enjoyed a little update on some of the upcoming features for OpenDroneMap. In addition to the above, we’ll also be wrapping in reporting and robustness improvements. More on that soon, as that is another huge piece that will help the entire community of users.

(This post CC BY-SA 4.0 licensed)

(Shout out to Pau Gargallo Piracés of Mapillary for the technical aspects of this write up. He is not responsible for any of the mistakes, generalities, and distortions in the technical aspects. Those are all mine).

Posted in 3D, Docker, OpenDroneMap, OpenDroneMap, PDAL | Tagged: , , | 2 Comments »

Taking Slices from ~~LiDAR~~ OpenDroneMap data: Part X

Posted by smathermather on February 23, 2017

Part 10 of N… , wait. This is a lie. This post is actually about optical drone data, not LiDAR data. This is about next phase features fro OpenDroneMap — automated and semiautomation of the point clouds, creation of DTMs and other fun such stuff.

To date, we’ve only extracted Digital Surface Models from ODM — the top surface of everything in the scene. As it is useful for hydrological modeling and other purposes to have a Digital Terrain Model estimated, we’ll be including PDAL’s Progressive Morphological Filter for the sake of DEM extraction. Here’s a small preview:

Posted in 3D, Docker, OpenDroneMap, PDAL | Tagged: , , | Leave a Comment »

Taking Slices from LiDAR data: Part IX

Posted by smathermather on February 20, 2017

Part 9 of N… , see e.g. my previous post on the topic.

We’ve been working to reduce the effect of overlapping samples on statistics we run on LiDAR data, and to do so, we’ve been using PDAL’s filters.sample approach. One catch: this handles the horizontal sampling problem well, but we might want to intentionally retain samples from high locations — after all, I want to see the trees for the forest and vice versa. So, it might behoove us to sample within each of our desired height classes to retain as much vertical information as possible.

Posted in 3D, Database, Docker, LiDAR, Other, PDAL, pointcloud, PostGIS, PostgreSQL | Tagged: , , , , , | 1 Comment »

Taking Slices from LiDAR data: Part VIII

Posted by smathermather on February 18, 2017

Part 8 of N… , see e.g. my previous post on the topic.

I didn’t think my explanation of sampling problems with LiDAR data in my previous post was adequate. Here are a couple more figures for clarification.

We can take this dataset over trees, water, fences, and buildings that is heavily sampled in some areas and sparsely sampled in others and use PDAL’s filters.sample (Poisson dart-throwing) to create an evenly sampled version of the dataset.

Figure showing overlap of LiDAR scanlines

Figure showing overlap of LiDAR scanlines

Figure showing data resampled for eveness

Figure showing data resampled for evenness

An extra special thanks to the PDAL team for not only building such cool software, but being so responsive to questions!

Posted in 3D, Database, Docker, LiDAR, Other, PDAL, pointcloud, PostGIS, PostgreSQL | Tagged: , , , , , | 1 Comment »

Taking Slices from LiDAR data: Part VII

Posted by smathermather on February 15, 2017

Part 7 of N… , see e.g. my previous post on the topic.

More work on taking LiDAR slices. This time, the blog post is all about data preparation. LiDAR data, in its raw form, often has scan line effects when we look at density of points.


This can affect statistics we are running, as our sampling effort is not even. To ameliorate this affect a bit, we can decimate our point cloud before doing further work with it. In PDAL, we have three choices for decimation: filters.decimation, which samples every Nth point from the point cloud; filters.voxelgrid, which does volumetric pixel based resampling; and filters.sample or “Poisson sampling via ‘Dart Throwing'”.

filters.decimation won’t help us with the above problem. Voxelgrid sampling could help, but it’s very regular, so I reject this on beauty grounds alone. This leaves filters.sample.

The nice thing about both the voxelgrid and the poisson sampling is that they retain much of the shape of the point cloud while down sampling the data:



We will execute the poisson sampling in PDAL. As many things in PDAL are best done with a (json) pipeline file, we construct a pipeline file describing the filtering we want to do, and then call that from the command line:

We can slice our data up similar to previous posts, and then look at the point density per slice. R-code for doing this forthcoming (thanks to Chris Tracey at Western Pennsylvania Conservancy and the LidR project), but below is a graphic as a teaser. For the record, we will probably pursue a fully PDAL solution in the end, but really interesting results in the interim:


More to come. Stay tuned.

Posted in 3D, Database, Docker, LiDAR, Other, PDAL, pointcloud, PostGIS, PostgreSQL | Tagged: , , , , , | 2 Comments »

Finding peace, finding ground: Drone flights for hydrologic modeling

Posted by smathermather on October 15, 2016


Lately I’ve been helping with project to predict flooding in Dar es Salaam (“Home of Peace”), the capital former capital of Tanzania in East Africa. This is a really fun project for me, in part because most of the hard work is already done, but there still remain some tricky problems to solve.

This project is a part of Dar Ramani Huria, a “community-based mapping project… training university students and local community members to create highly accurate maps of the most flood-prone areas of the city using OpenStreetMap.”

Dar es Salaam (or Dar for short), is the fastest growing city in Africa, making the identification of flood prone zones during rainy seasons absolutely critical. The Ramani Huria crew has mapped the streams, ditches, and other waterways of Dar, as well as flown imagery (via drone) over much of the city critical parts of the city.


Problem Space:

The results are stunning, but using drone imagery and photogrammetric point clouds (instead of LiDAR) has it’s limitations.

One problem, that I’m not going to get into in any depth today, is the difficulty of conflating (vertically and horizontally) different datasets flown at different times with commodity GPS.

Another problem is the difficulty of turning photogrammetrically derived point clouds into Digital Terrain Models. There is proprietary software that does this well (e.g. LasTools and others), but we seek a free and open source alternative. Let’s visualize the problem. See below a photogrammetrically-derived point cloud (processed in Pix4D, visualized at


We can directly turn this into a digital surface model (DSM):



We can see the underlying topography, but buildings and trees are much of the signal in the surface model. If we, for example, calculate height above the nearest stream as a proxy for flooding, we get a noisy result.

OSM to the Rescue

If you recall at the beginning of this post, I said this is “a really fun project for me” because “most of the hard work is already done”. Dar is a city with pretty much every building digitized. We can take advantage of this work to remove buildings from our point cloud. We’ll use the utilities associated with the Point Data Abstraction Layer (PDAL) library. Specifically, we’ll use PDAL’s ability to clip point cloud data with 2D OGR compatible geometries (a great tutorial on this here).

First thing we need to do is extract the buildings from OSM into shapefiles. For this, and easy way is to use Mapzen’s Metro Extracts:


I chose Datasets grouped into individual layers by OpenStreetMap tags (IMPOSM) as this gives me the capacity to just choose “Buildings”. I loaded those buildings into a PostGIS database for the next steps. What I seek is a shapefile of all the areas which are not buildings, essentially from this:


To this:


For this we’ll need to use ST_Difference + ST_ConvexHull. ST_ConvexHull will create a shape of the extent of our data, and we’ll use ST_Difference to cookie-cutter out all the building footprints:

DROP TABLE IF EXISTS "salaam-buff-diff-split1";
CREATE TABLE "salaam-buff-diff-split1" AS
WITH subset AS (
	SELECT ST_Transform(geom, 32737) AS geom FROM "salaam-buildings" LIMIT 20000
extent AS (
	SELECT ST_ConvexHull(ST_Union(geom)) AS geom FROM subset
unioned AS (
	SELECT ST_Union(geom) AS geom FROM subset
differenced AS (
	SELECT ST_Difference(a.geom, b.geom) AS geom FROM
	extent a, unioned b
SELECT 1 AS id, geom FROM differenced;

This gives us a reasonable result. But, when we use this in PDAL, (I assume) we want a subdivided shape to ensure we don’t have to access the entire point cloud at any given time to do our clipping. We’ll add to this process ST_SubDivide, which will subdivide our shape into parts not to exceed a certain number of nodes. In this case we’ll choose 500 nodes per shape:

DROP TABLE IF EXISTS "salaam-buff-diff-split1";
CREATE TABLE "salaam-buff-diff-split1" AS
WITH subset AS (
	SELECT ST_Transform(geom, 32737) AS geom FROM "salaam-buildings" LIMIT 20000
extent AS (
	SELECT ST_ConvexHull(ST_Union(geom)) AS geom FROM subset
unioned AS (
	SELECT ST_Union(geom) AS geom FROM subset
differenced AS (
	SELECT ST_Difference(a.geom, b.geom) AS geom FROM
	extent a, unioned b
SELECT 1 AS id, ST_Subdivide(geom, 500) AS geom FROM


Finally, if we want to be sure to remove the points from the edges of buildings (we can assume the digitized buildings won’t perfectly match our point clouds), then we should buffer our shapes:

DROP TABLE IF EXISTS "salaam-buff-diff-split1";
CREATE TABLE "salaam-buff-diff-split1" AS
WITH subset AS (
	SELECT ST_Transform(geom, 32737) AS geom FROM "salaam-buildings" LIMIT 20000
buffered AS (
	SELECT ST_Buffer(geom, 2, 'join=mitre mitre_limit=5.0') AS geom FROM subset
extent AS (
	SELECT ST_ConvexHull(ST_Union(geom)) AS geom FROM subset
unioned AS (
	SELECT ST_Union(geom) AS geom FROM Buffered
differenced AS (
	SELECT ST_Difference(a.geom, b.geom) AS geom FROM
	extent a, unioned b
SELECT 1 AS id, ST_Subdivide(geom, 500) AS geom FROM

If we recall our before point cloud:



Now we have filtered out most buildings:



In order to do this filtering in PDAL, we do two things. First we create a json file that defines the filter:


Then we use these json definitions to apply the filter:

nohup sudo docker run -v /home/gisuser/docker/:/data pdal/pdal:1.3 pdal pipeline /data/shape-clip.json

Problems continue

This looks pretty good, but as we interrogate the data, we can see the artifacts of trees and some buildings still linger in the dataset.




Enter the Progressive Morphological Filter

It is possible to use the shape of the surface model to filter out buildings and trees. To do so, we start with the assumption that the buildings and vegetation shapes distinctive from the shape of the underlying ground. We have already used the hard work of the OpenStreetMap community to filter most of the buildings, but we still have some buildings and plenty of trees. PDAL has another great tutorial for applying this filter which we’ll leverage.

Again we need a JSON file to define our filter:

  "pipeline": {
    "name": "Progressive Morphological Filter with Outlier Removal",
    "version": 1.0,
    "filters": [{
        "name": "StatisticalOutlierRemoval",
        "setMeanK": 8,
        "setStddevMulThresh": 3.0
      }, {
        "name": "ProgressiveMorphologicalFilter",
        "setCellSize": 1.5

And then use that filter to remove all the morphology and statistical outliers we don’t want:

nohup sudo docker run -v /home/gisuser/docker/:/data pdal/pdal:1.3 pdal pcl -i /data/2015-05-20_tandale_merged_densified_point_cloud_part_1_nobuild_buff.las -o /data/nobuild_filtered2.las -p /data/sor-pmf.json

This command will remove our colorization for the points, so we’ll see the colorization according to height only:




What remains is to interpolate this into digital terrain model. That is a project for another day.

Posted in 3D, Drone, Other, PDAL, Photogrammetry, UAS | Tagged: , , , | 1 Comment »

Viewing Sparse Point Clouds from OpenDroneMap — GeoKota

Posted by smathermather on June 30, 2016

This is a post about OpenDroneMap, an opensource project I am a maintainer for. ODM is a toolchain for post-processing drone imagery to create 3D and mapping products. It’s currently in beta and under pretty heavy development. If you’re interested in contributing to the project head over here. The Problem So for most of the […]

via Viewing Sparse Point Clouds from OpenDroneMap — GeoKota

Posted in 3D, Image Processing, OpenDroneMap, OpenDroneMap, Optics, Other, Photogrammetry | Tagged: , , , | Leave a Comment »