forked from dicengine/dicegui
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcal_help.html
81 lines (81 loc) · 5.07 KB
/
cal_help.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
<!DOCTYPE html>
<html>
<head>
<meta charset="UTF-8">
<title>Calibration Help</title>
<script>
if (typeof module === 'object') {
window.module = module;
module = undefined;
}
</script>
<script src="./js/jquery.min.js"></script>
<script src="./js/jquery-ui.js"></script>
<script>
if (window.module)
module = window.module;
</script>
<script type="text/javascript" src="./js/loader.js"></script>
<script src="./js/requires.js" defer></script>
<!-- Local style sheet -->
<link rel="stylesheet" type="text/css" href="style.css">
<!-- Google fonts and icons -->
<link rel="stylesheet" type="text/css" href="open_sans.css">
<link rel="stylesheet" type="text/css" href="material_icons.css">
</head>
<body style="background-color: white;">
<div style="background-color: white;">
<div style="padding: 20px;">
<h2>Performing a Stereo or Single Camera Calibration</h2>
<p>Select 'perform cal' to run a calibration given a set of images using the OpenCV
calibration tools through DICe.</p>
<p>The first step in performing a calibration is selecting the image file folder. Once
this folder is selected, DICe will attempt to automatically determine the file naming
convention. If this fails, the components of the file naming convention need to be filled in
manually. If the components of the naming convention are filled in correctly, an example
image name will appear in green at the bottom of the first column on the calibration tool.
If the name appears in red, this indicates that the file pattern is invalid. (note: all numbers
in the filenames should be padded with zeros, for example if there are 10 images the image numbering
should be _00 ... _10, not _0 ... _10, or for 100 images _000 ... _100, not _0 ... _100.)</p>
<p>The second step is to select the pattern parameters, beginning with the type of
calibration plate that is used in the images. There are three options for this:
checkerboard, black dots on a white background, and white dots on a black background. Once
the pattern type is selected, the user must input the dimensions of the board, including the
spacing size.</p>
<p>To preview the calibration marker extraction for a particular image, click the
'Preview' button in the left column of the tool. If the marker extraction is successful, the
images will appear with a green border. A red border indicates a failed extraction.</p>
<p>For checkerboard patterns, there are thresholding options that can be adjusted. For dot
patterns, the default thresholding used involves stepping through a series of constant
binary threshold values until the marker dots can be extracted. The threshold setting can be
changed by de-selecting the auto-threshold check box and setting the parameters.</p>
<p>If the calibration is successful, a cal.txt file will be created in the working
directory. This file must then be loaded by pressing the 'accept' button in the calibration
tool. An existing calibration file can be loaded using the 'load cal' menu option and
selecting the file.</p>
<p>For a stereo calibration, the epipolar error should be less than 0.5 pixels. Large
epipolar errors happen when one or more of the images introduces too much reprojection
error. If this occurrs, look at the cal.log file in the working directory to see which
images have the highest error. Through trial and error, eliminate images by selecting them
in the list in the right column and re-running the calibration until the error is less than 0.5 pixels.</p>
<p>For single camera calibration, only the intrinsic parameters are estimated (such as the
focal lengths, image centroids, etc). No reprojection error is calculated so there isn't a
quality metric for the single camera calibration other than the RMS error.</p>
<h2>Loading an Existing Calibration File</h2>
<p>If the calibration has already been completed, or you would like to use an existing
calibration file, select the 'load cal' menu option and select the desired file. DICe uses a
specific xml format for calibration files, but can also read VIC3D xml files.</p>
<h1>VIC3D XML Format</h1>
<p>This is a format similar to the one used in a VIC3D .z3d file. Note that the
orientations for each camera are not with respect to each other, but the origin between the
two cameras. The file extension must be .xml for this format to be used (.z3d files must be
unzipped. The .xml file from the archive can then be loaded in DICe)</p>
<xmp> <calibration lri="calibration"> <camera id="0">182.694
184.935 2428.49 2431.72 3.39496 0.0881161 0 0 <orientation>-179.925 -14.7548
0.578387 1644 5.43174 970.919</orientation> </camera> <camera id="1">211.986 191.741 2479.9 2484.28 2.99616
0.112661 0 0 <orientation>179.77 14.7532 -0.617865 -1765.32 -49.9461 1075.75</orientation> </camera> </calibration> </xmp>
</div>
</div>
<!-- end content-->
</body>
</html>