How to get remained points after using RANSAC Shape Detection plugin

[JamesRayen]

Hi,I'm using the plugin RANSAC Shape Detection to detect planes,it can detect planes with high efficiency,but I want to get those ponits which are not detected by the plugin,but I don't know how to do.Is there anyone who can help me?THANK YOU!

[daniel]

You can merge all the subsests of points corresponding to each plane. These subsets are exported in the 'Ransac Detected Shapes' group. You can select all of them and then merge them (Edit > Merge).

Then you can compute the distances between the merged cloud and the original one (use the merged cloud as 'reference'). And eventually extract the points from the original cloud with non zero distances (with Edit > Scalar Fields > Filter by value).

[JamesRayen]

You can merge all the subsests of points corresponding to each plane. These subsets are exported in the 'Ransac Detected Shapes' group. You can select all of them and then merge them (Edit > Merge).

Then you can compute the distances between the merged cloud and the original one (use the merged cloud as 'reference'). And eventually extract the points from the original cloud with non zero distances (with Edit > Scalar Fields > Filter by value).

hello,daniel, I do as you said, but in the last step I find that Filter by value is gray and I can't toggle it.I have tried to select the merged plane or the original cloud data, but it didnot work.I don't know where I am wrong,so could you kindly tell me in detail?

[daniel]

Have you actually computed the distances? (you have to hit the 'Compute' button).

As the reference should be the merged cloud, the distances scalar field should be created on the original cloud. Therefore you have to select the original cloud to apply the Filter by Value tool.

[JamesRayen]

Have you actually computed the distances? (you have to hit the 'Compute' button).

As the reference should be the merged cloud, the distances scalar field should be created on the original cloud. Therefore you have to select the original cloud to apply the Filter by Value tool.

Thank you,daniel,I followed your advise and made it.But I found it was a littile different of point number.There are 3404 points in my dataset, and the plugin extracted 2941 points.After I used Filter by Value tool,I found the extracted points is 2937.And then, I used another two datasets, but there was still 4 points which were not in the extracted points by Filter by Value tool.Is this a debug?Or something I did wrong?I have tried to set the range of Filter by Value as 0-0 and 0-0.001,0-0.005,but the results were the same.
Another question is how to paste image in the forume?THANKS for your kindly reply!

[daniel]

To paste an image, use the 'Attachments' tab below the message form (then click on the 'Add files' button, choose a file, and once it is loaded, click on the 'place inline' button next to the image file - it will add a special block in the message text where your cursor is = where the image will appear). Pay attention that the size of the image you can import is limited (= you may have to resize the image first).

And for the missing number of points: on my side, when I compute the distances, and afterwards extract the points in the [0;0] range, I get the exact same number of points as in the merged cloud. Don't hesitate to send me your cloud if you want me to check this with your data (cloudcompare [at] danielgm.net).

[JamesRayen]

To paste an image, use the 'Attachments' tab below the message form (then click on the 'Add files' button, choose a file, and once it is loaded, click on the 'place inline' button next to the image file - it will add a special block in the message text where your cursor is = where the image will appear). Pay attention that the size of the image you can import is limited (= you may have to resize the image first).

And for the missing number of points: on my side, when I compute the distances, and afterwards extract the points in the [0;0] range, I get the exact same number of points as in the merged cloud. Don't hesitate to send me your cloud if you want me to check this with your data (cloudcompare [at] danielgm.net).

Thank you,daniel,I have sent you my test dataset.