Commit f160d8c8 authored by Roddy Collins's avatar Roddy Collins

Merge branch '2018-03-camera-models' into 'master'

Add camera models for the March 2018 collection event.

See merge request !2
parents eda25f2f 70523f82
This directory contains various metadata regarding the MEVA data.
Subirectories include:
[camera-models](camera-models): Camera models for the MEVA KF1 collection.
This directory contains the geometric scene model for the March 2018
MEVA data collection. The format of the data is a common coordinate
system containing a calibrated camera model for each camera and a
collection of 3D landmark points with their annotated 2D locations in
each image that sees them. Additionally the model is georegistered.
This directory contains models covering a subset of the public
cameras. We started with the stationary (non-patrolling)
RGB cameras that are outdoors and have overlapping fields of view.
In the future we will add models for IR cameras and each state of
the patrolling cameras. We may also provide model of indoor cameras,
but these will likely be in different local coordinates that are
not georegistered.
The data is structured as follows:
* [cameras](cameras) is a directory of camera files, one per camera
id/state. Refer to the [cameras readme](cameras/README.md)
for details on the format.
* [frames](frames) is a directory of sample images, one per camera
id/state. The file names match the camera names, but with a different
file extension. These files are mostly here for convenience to help
visualize the geometry in the image space without needing the video.
* [frames.txt](frames.txt) is an ordered list of the frames (and
corresponding cameras) that are currently considered for the model.
The order of the image name in this file assigned it an integer index.
The first line in this file is camera 1, the second is 2, and so on.
* [tracks.txt](tracks.txt) is a text file containing the image location
of selected landmark points in each image. Each line provides a the
following `track_id camera_id X_loc Y_loc`. There are several other
fields that appear on each line as well, but these are currently not
used for this application and appear as default values
`1 1 0 255 255 255 0`. The `camera_id` indicates the camera index
that sees the current landmark, it match the line number in `frames.txt`.
* [landmarks.ply](landmarks.ply) is sparse 3D point cloud of selected
landmark points. PLY is a standard ASCII file format for 3D mesh
and point data. The header documents the order of the fields. The
relevant fields are the first three, `X Y Z` location, and the last
two `track_id observations`. The `track_id` is a unique number for
each track/landmark and matches the `track_id` in `tracks.txt`. The
`observations` field indicates the number of cameras that see the
landmark. You can open PLY files in many 3D applications such as
[MeshLab](http://www.meshlab.net/) and [Blender](https://www.blender.org).
* [gcps.json](gcps.json) is a geo-JSON formatted file that contains
a few manually selected ground control points (GCPs) that were used
to georegister the coordinate system. You should not need to use
these GCPs directly, see `geo_origin.txt`.
* [geo_origin.txt](geo_origin.txt) a simple ASCII file containing only
three numbers: `latitude longitude elevation` in WGS84. This is the
location of the origin of the local cartesian coordinate system in
geographic coordinates.
* [meva-2018-03-telesculptor.conf](meva-2018-03-telesculptor.conf)
This is an optional project configuration file that enables loading
all of the above data into the open source
[TeleSculptor](https://github.com/kitware/TeleSculptor) application
for visualization of the camera geometry. TeleSculptor, along with
other tools, was used to build this model. If you have all of the data
checked out and have TeleSculptor v1.0 you can use
`File -> Open Project` and select this file to view the data.
* [homographies](homographies) (coming soon) contains alternative models for direct
image-to-image mapping with a homography transformation. These models
are only applied in cases of co-located RGB and IR cameras that have similar viewpoints and minimal parallax.
1460.19079702 0 959.5
0 1460.19079702 535.5
0 0 1
-0.882264787742 -0.470041487771 0.0258813462527
-0.020547611134 -0.0164750029126 -0.999653124816
0.470304837298 -0.882490551821 0.00487708524067
127.469270361 12.7661662918 98.4895022221
-0.353997555311 0.122164890204
2196.4170273 0 959.5
0 2196.4170273 539.5
0 0 1
-0.997424546078 0.0677089316475 0.0236595742968
-0.0111311709675 0.179743630005 -0.98365050933
-0.0708545828697 -0.981380521535 -0.178527028874
-0.57076048718 -16.2977260539 144.709029621
-0.188099076566 0
2997.58983856 0 959.5
0 2997.58983856 539.5
0 0 1
0.520809766882 -0.853382581137 0.0222566154673
-0.130140267522 -0.105136016443 -0.9859056389
0.843694670772 0.510572804073 -0.165815301623
-11.2857171381 0.81503879573 15.4308860251
-0.188099076566 0
3361.20904041 0 960
0 3361.20904041 540
0 0 1
0.483680462594 0.875017230691 0.0199513432618
0.202319394696 -0.0895995448253 -0.975212173887
-0.851539824456 0.475727619085 -0.220370505751
27.9907720829 0.598570439881 61.5135953027
0 0
4004.96723614 0 959.5
0 4004.96723614 539.5
0 0 1
-0.924718286681 0.380641188918 -0.00289405923643
0.0509501255898 0.116235068867 -0.991914055485
-0.377226954209 -0.917388518603 -0.126878410115
25.7075161779 -11.6966747468 171.258160563
0 0
3751.8219329 0 959.5
0 3751.8219329 539.5
0 0 1
0.983236711091 0.180664407665 -0.0246158844372
0.00632924370911 -0.168741296753 -0.985640053693
-0.182223792732 0.968961684781 -0.16705610671
-14.7137926596 1.21671519903 75.7285808212
-0.188099076566 0
1452.84910798 0 959.5
0 1452.84910798 535.5
0 0 1
-0.997592340867 -0.0532273874405 0.0444563456564
-0.0595258939458 0.328278929156 -0.942703353459
0.0355835550898 -0.943079948837 -0.330656953214
-15.3004829927 10.9596122579 -21.6448926086
-0.353997555311 0.122164890204
1490.5193438 0 959.5
0 1490.5193438 535.5
0 0 1
0.992650918786 0.0959763451094 -0.073706815233
-0.0339568246321 -0.363696666464 -0.930898312849
-0.116151140732 0.926559914945 -0.357764778205
17.4227724787 -9.5785787824 31.8484366523
-0.353997555311 0.122164890204
This directory contains camera calibration (intrinsic and extrinsic)
for the March 2018 MEVA data collection. The goal is to provide camera
models in a common coordinate system to allow tools to project points
from one camera into the world space and then into other cameras. The
local camera coordinate system is scaled to units of meters. The Z=0
plane is roughly aligned with the ground. This is an ENU (East, North,
Up) coordinate system. The geo_origin.txt file in the parent directory
provides the geographic location of the origin of the cooridinate system.
The data here is stored in KRTD format.
KRTD is an ASCII file containing the following matrices:
K - 3x3 upper triangular intrinisic matrix encoding focal length,
principal point, skew, pixel aspect ratio.
R - 3x3 orthonormal rotation matrix
T - 3d translation vector
D - Nd vector of lens distortion parameters (using OpenCV conventions)
frames/G328.png
frames/G336.png
frames/G340.png
frames/G341.png
frames/G424.png
frames/G436.png
frames/G505.png
frames/G506.png
frames/G509.png
frames/G638.png
frames/G639.png
{
"features": [
{
"geometry": {
"coordinates": [
-85.529558,
39.049284,
243
],
"type": "Point"
},
"properties": {
"location": [
-20.775880255474576,
5.081582617241264,
0.3593906054064351
],
"name": "Subway Sidewalk",
"userRegistered": true
},
"type": "Feature"
},
{
"geometry": {
"coordinates": [
-85.529265,
39.04907,
243
],
"type": "Point"
},
"properties": {
"location": [
5.4472233337265745,
-19.10106825432819,
0.7207363250042702
],
"name": "Bus Station Light",
"userRegistered": true
},
"type": "Feature"
},
{
"geometry": {
"coordinates": [
-85.52876,
39.049242,
243
],
"type": "Point"
},
"properties": {
"location": [
48.88530485077557,
1.402794257351851,
0.7110465578396865
],
"name": "Trailer Light",
"userRegistered": true
},
"type": "Feature"
},
{
"geometry": {
"coordinates": [
-85.528496,
39.050078,
243
],
"type": "Point"
},
"properties": {
"location": [
70.90750975929316,
94.80040584339383,
0.44586060245302495
],
"name": "Stop Sign School Parking",
"userRegistered": true
},
"type": "Feature"
},
{
"geometry": {
"coordinates": [
-85.529243,
39.050116,
242
],
"type": "Point"
},
"properties": {
"location": [
5.334435608236992,
95.94855420745917,
-0.18420633263722874
],
"name": "School Outer Circle Light",
"userRegistered": true
},
"type": "Feature"
}
],
"type": "FeatureCollection"
}
39.0491071345 -85.529451881 242.348566139
\ No newline at end of file
ply
format ascii 1.0
comment written by VITAL
element vertex 82
property float x
property float y
property float z
property float nx
property float ny
property float nz
property uchar red
property uchar green
property uchar blue
property uint track_id
property uint observations
end_header
73.2599 115.363 1.39131 0 0 0 255 255 255 0 2
69.9478 114.204 1.7912 0 0 0 255 255 255 1 2
70.8755 94.9027 0.460621 0 0 0 255 255 255 2 2
77.3021 92.7984 0.812612 0 0 0 255 255 255 3 2
74.6511 93.4463 0.745902 0 0 0 255 255 255 4 2
69.5378 36.6767 1.21137 0 0 0 255 255 255 5 2
74.5271 72.13 1.39512 0 0 0 255 255 255 6 3
78.863 89.9104 0.766267 0 0 0 255 255 255 7 2
80.8305 115.662 2.39816 0 0 0 255 255 255 8 2
80.4458 117.237 2.3762 0 0 0 255 255 255 9 2
75.5469 112.338 2.25593 0 0 0 255 255 255 10 2
75.7763 112.733 0.853545 0 0 0 255 255 255 11 2
76.5259 87.0752 2.10747 0 0 0 255 255 255 12 2
-13.9441 -9.07794 0.213253 0 0 0 255 255 255 14 3
39.742 41.9825 1.32346 0 0 0 255 255 255 15 2
29.6851 41.6724 1.2432 0 0 0 255 255 255 16 2
-13.9706 -9.22131 4.8246 0 0 0 255 255 255 17 3
-11.6302 -13.5895 2.66211 0 0 0 255 255 255 18 3
-7.61253 -35.2791 3.8176 0 0 0 255 255 255 19 3
-21.6848 -10.8382 2.3514 0 0 0 255 255 255 21 2
-21.1328 -10.7381 1.40367 0 0 0 255 255 255 22 2
59.0481 -6.46366 3.39576 0 0 0 255 255 255 23 2
48.9015 1.41142 0.769735 0 0 0 255 255 255 24 3
28.3176 -0.267457 0.81846 0 0 0 255 255 255 25 2
-18.1274 -8.75249 2.81406 0 0 0 255 255 255 27 2
1.72543 19.823 0.564346 0 0 0 255 255 255 29 4
5.46524 -18.9937 0.683204 0 0 0 255 255 255 30 3
-1.37732 -7.29236 0.517425 0 0 0 255 255 255 31 2
-0.633556 -10.0839 0.551258 0 0 0 255 255 255 32 2
-7.55966 -35.3077 5.31693 0 0 0 255 255 255 33 3
-11.1788 -34.4287 -0.0408538 0 0 0 255 255 255 34 3
-12.4053 -29.6329 -0.00913024 0 0 0 255 255 255 35 2
-13.5777 -24.8223 0.0490198 0 0 0 255 255 255 36 2
-9.08011 -22.8809 0.00616277 0 0 0 255 255 255 37 2
-10.7627 -22.2368 5.53071 0 0 0 255 255 255 38 2
-10.7651 -22.2536 3.98956 0 0 0 255 255 255 39 2
14.3652 -23.3905 0.750885 0 0 0 255 255 255 41 3
25.4183 3.12477 0.517186 0 0 0 255 255 255 42 2
-10.9101 -21.7474 0.0611733 0 0 0 255 255 255 44 2
-13.5425 -24.8482 1.14246 0 0 0 255 255 255 45 2
-12.3728 -29.63 1.10558 0 0 0 255 255 255 46 3
-11.1854 -34.4199 1.05605 0 0 0 255 255 255 47 4
-14.7856 -30.1987 1.10946 0 0 0 255 255 255 48 3
-14.9912 -29.3787 1.1165 0 0 0 255 255 255 49 3
-15.6365 -30.3929 1.10981 0 0 0 255 255 255 50 3
-13.2195 -29.8381 1.10665 0 0 0 255 255 255 51 3
-14.0404 -70.3204 2.20603 0 0 0 255 255 255 52 2
-38.577 -91.627 1.28387 0 0 0 255 255 255 53 2
-19.9153 -54.3718 0.962248 0 0 0 255 255 255 54 2
-13.6164 -34.9815 1.07193 0 0 0 255 255 255 55 2
4.95937 59.8405 0.169688 0 0 0 255 255 255 57 2
5.23043 97.2976 -0.266385 0 0 0 255 255 255 58 2
-15.9575 -25.4121 1.15591 0 0 0 255 255 255 64 2
-15.9706 -25.3676 0.0318477 0 0 0 255 255 255 65 2
-14.3793 -25.0341 1.15067 0 0 0 255 255 255 66 2
-14.3947 -25.0009 0.0358745 0 0 0 255 255 255 67 2
-12.6014 -22.0562 1.15408 0 0 0 255 255 255 68 2
-13.2766 -22.2244 2.19233 0 0 0 255 255 255 69 2
-14.9504 -22.6077 2.09206 0 0 0 255 255 255 70 2
-16.5511 -22.9628 2.0945 0 0 0 255 255 255 71 2
-14.4739 -35.1794 1.06704 0 0 0 255 255 255 72 2
3.45548 -60.7821 0.301878 0 0 0 255 255 255 73 2
27.8275 0.10445 0.892741 0 0 0 255 255 255 74 2
8.64353 -3.4533 0.485162 0 0 0 255 255 255 75 2
97.3119 67.0473 3.1308 0 0 0 255 255 255 81 2
96.9833 69.3486 3.13771 0 0 0 255 255 255 83 2
70.9637 94.6398 3.28712 0 0 0 255 255 255 87 2
74.5374 72.1588 0.464529 0 0 0 255 255 255 88 3
16.0432 106.148 -0.232758 0 0 0 255 255 255 96 2
-10.9605 -21.7287 4.63899 0 0 0 255 255 255 105 2
8.32138 -2.30554 0.494653 0 0 0 255 255 255 129 2
-11.408 -33.5992 -0.0319515 0 0 0 255 255 255 131 4
-11.3938 -33.5829 1.06042 0 0 0 255 255 255 132 4
-12.584 -28.7955 1.1154 0 0 0 255 255 255 133 3
-13.7474 -23.9984 1.13649 0 0 0 255 255 255 135 2
-12.4219 -15.8349 1.49097 0 0 0 255 255 255 138 2
-12.6298 -14.9529 1.48607 0 0 0 255 255 255 139 2
-11.8137 -18.1483 1.22479 0 0 0 255 255 255 140 3
-13.0988 -12.779 1.29837 0 0 0 255 255 255 141 3
-11.8358 -18.1254 2.20483 0 0 0 255 255 255 143 3
-10.7122 -35.1293 -0.104521 0 0 0 255 255 255 146 3
28.7895 39.6295 1.33726 0 0 0 255 255 255 154 2
ROI = 1e+299 1e+299 1e+299 -1e+299 -1e+299 -1e+299
geo_origin_file = geo_origin.txt
ground_control_points_file = gcps.json
output_krtd_dir = cameras
output_ply_file = landmarks.ply
output_tracks_file = tracks.txt
video_reader:image_list:image_reader:type = ocv
video_reader:type = image_list
video_source = frames.txt
This diff is collapsed.
This directory contains camera models for the MEVA KF1 collection.
The KF1 collection took place in two separate events, in March and May of 2018. Between the events, camera positions may have shifted; for the moment the camera models are broken out by collection month. As we refine our models, this organization may change.
This directory includes:
* [2018-03](2018-03) is a directory containing models for the March 2018 collection event.
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment