<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Noisy depth data in Software Archive</title>
    <link>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055047#M51469</link>
    <description>&lt;P&gt;Hi,&lt;/P&gt;

&lt;P&gt;I am using the R200 to capture depth images that I convert to point clouds. An issue I'm finding is that planar surfaces can be extremely noisy.&lt;/P&gt;

&lt;P&gt;Settings:&lt;/P&gt;

&lt;PRE class="brush:plain;"&gt;Gain: 2
Exposure: 30ms
Resolution: 628x468&lt;/PRE&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;&lt;span class="lia-inline-image-display-wrapper" image-alt="wallNoise.PNG"&gt;&lt;img src="https://community.intel.com/t5/image/serverpage/image-id/8191i3028E948DD7678FB/image-size/large?v=v2&amp;amp;px=999&amp;amp;whitelist-exif-data=Orientation%2CResolution%2COriginalDefaultFinalSize%2CCopyright" role="button" title="wallNoise.PNG" alt="wallNoise.PNG" /&gt;&lt;/span&gt;&lt;/P&gt;

&lt;P&gt;My first question is whether I can reduce this via my capturing method? (Note I would use a longer exposure if I could but it does not seem to let me without crashing). If not and my settings are fine would it be possible to smooth the surfaces down well enough that they look nice?&lt;/P&gt;

&lt;P&gt;I have tried smoothing the PXCImage myself before creating the projection to vertices as well as some filtering of the clouds later but without much luck. The results are still reasonably noisy.&lt;/P&gt;

&lt;P&gt;Any advice? Thanks.&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
    <pubDate>Tue, 10 Nov 2015 09:57:05 GMT</pubDate>
    <dc:creator>Thomas_5207</dc:creator>
    <dc:date>2015-11-10T09:57:05Z</dc:date>
    <item>
      <title>Noisy depth data</title>
      <link>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055047#M51469</link>
      <description>&lt;P&gt;Hi,&lt;/P&gt;

&lt;P&gt;I am using the R200 to capture depth images that I convert to point clouds. An issue I'm finding is that planar surfaces can be extremely noisy.&lt;/P&gt;

&lt;P&gt;Settings:&lt;/P&gt;

&lt;PRE class="brush:plain;"&gt;Gain: 2
Exposure: 30ms
Resolution: 628x468&lt;/PRE&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;&lt;span class="lia-inline-image-display-wrapper" image-alt="wallNoise.PNG"&gt;&lt;img src="https://community.intel.com/t5/image/serverpage/image-id/8191i3028E948DD7678FB/image-size/large?v=v2&amp;amp;px=999&amp;amp;whitelist-exif-data=Orientation%2CResolution%2COriginalDefaultFinalSize%2CCopyright" role="button" title="wallNoise.PNG" alt="wallNoise.PNG" /&gt;&lt;/span&gt;&lt;/P&gt;

&lt;P&gt;My first question is whether I can reduce this via my capturing method? (Note I would use a longer exposure if I could but it does not seem to let me without crashing). If not and my settings are fine would it be possible to smooth the surfaces down well enough that they look nice?&lt;/P&gt;

&lt;P&gt;I have tried smoothing the PXCImage myself before creating the projection to vertices as well as some filtering of the clouds later but without much luck. The results are still reasonably noisy.&lt;/P&gt;

&lt;P&gt;Any advice? Thanks.&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Tue, 10 Nov 2015 09:57:05 GMT</pubDate>
      <guid>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055047#M51469</guid>
      <dc:creator>Thomas_5207</dc:creator>
      <dc:date>2015-11-10T09:57:05Z</dc:date>
    </item>
    <item>
      <title>That is just the reality of</title>
      <link>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055048#M51470</link>
      <description>&lt;P&gt;That is just the reality of 3D. It is hard, and sensors are noisy. Our brain does an amazing job that we take for granted, but computers are different to brains...&lt;/P&gt;

&lt;P&gt;You could use a third party library. Have a look at PCL for a starting point on filters, segmentation, reconstruction, etc...&lt;/P&gt;</description>
      <pubDate>Wed, 11 Nov 2015 01:10:48 GMT</pubDate>
      <guid>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055048#M51470</guid>
      <dc:creator>samontab</dc:creator>
      <dc:date>2015-11-11T01:10:48Z</dc:date>
    </item>
    <item>
      <title>Also, grab points directly</title>
      <link>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055049#M51471</link>
      <description>&lt;P&gt;Also, grab points directly using a grabber.&lt;/P&gt;

&lt;P&gt;This should get you started :&amp;nbsp;&lt;A href="https://github.com/taketwo/rs"&gt;https://github.com/taketwo/rs&lt;/A&gt;&lt;/P&gt;</description>
      <pubDate>Wed, 11 Nov 2015 03:27:52 GMT</pubDate>
      <guid>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055049#M51471</guid>
      <dc:creator>Ashwin_S_1</dc:creator>
      <dc:date>2015-11-11T03:27:52Z</dc:date>
    </item>
    <item>
      <title>Quote:Ashwin S. wrote:</title>
      <link>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055050#M51472</link>
      <description>&lt;P&gt;&lt;/P&gt;&lt;BLOCKQUOTE&gt;Ashwin S. wrote:&lt;BR /&gt;&lt;P&gt;&lt;/P&gt;

&lt;P&gt;Also, grab points directly using a grabber.&lt;/P&gt;

&lt;P&gt;This should get you started :&amp;nbsp;&lt;A href="https://github.com/taketwo/rs" rel="nofollow"&gt;https://github.com/taketwo/rs&lt;/A&gt;&lt;/P&gt;

&lt;P&gt;&lt;/P&gt;&lt;/BLOCKQUOTE&gt;&lt;P&gt;&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;I appreciate the code to look at. What is the benefit of using this grabber over what I am doing presently? I can already take the depth stream and query the vertices to get the projection which I construct a point cloud from.&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Wed, 11 Nov 2015 10:06:47 GMT</pubDate>
      <guid>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055050#M51472</guid>
      <dc:creator>Thomas_5207</dc:creator>
      <dc:date>2015-11-11T10:06:47Z</dc:date>
    </item>
    <item>
      <title>Oh,</title>
      <link>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055051#M51473</link>
      <description>&lt;P&gt;Oh,&lt;BR /&gt;
	I should've mentioned from a workflow standpoint, if you were using PCL this would be helpful.&lt;BR /&gt;
	It also features different capturing modes and so on.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;No real benefit apart from that, since you have the depth to pointcloud stuff already working.&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Wed, 11 Nov 2015 19:46:57 GMT</pubDate>
      <guid>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055051#M51473</guid>
      <dc:creator>Ashwin_S_1</dc:creator>
      <dc:date>2015-11-11T19:46:57Z</dc:date>
    </item>
    <item>
      <title>Quote:Ashwin S. wrote:</title>
      <link>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055052#M51474</link>
      <description>&lt;P&gt;&lt;/P&gt;&lt;BLOCKQUOTE&gt;Ashwin S. wrote:&lt;BR /&gt;&lt;P&gt;&lt;/P&gt;

&lt;P&gt;Oh,&lt;BR /&gt;
	I should've mentioned from a workflow standpoint, if you were using PCL this would be helpful.&lt;BR /&gt;
	It also features different capturing modes and so on.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;No real benefit apart from that, since you have the depth to pointcloud stuff already working&lt;/P&gt;

&lt;P&gt;&lt;/P&gt;&lt;/BLOCKQUOTE&gt;&lt;P&gt;&lt;/P&gt;

&lt;P&gt;I am using PCL yes. The code was useful and gave me some ideas, thanks!&lt;/P&gt;</description>
      <pubDate>Thu, 12 Nov 2015 09:15:21 GMT</pubDate>
      <guid>https://community.intel.com/t5/Software-Archive/Noisy-depth-data/m-p/1055052#M51474</guid>
      <dc:creator>Thomas_5207</dc:creator>
      <dc:date>2015-11-12T09:15:21Z</dc:date>
    </item>
  </channel>
</rss>

