Good morning TBC’ers! Have you ever collected a mobile mapping dataset making a loop around a street block and noticed a gap in the middle of your point cloud data? And were you then unhappy with the automatic classification results in TBC? Well, today I will show you a handy classification trick in TBC to improve your mobile mapping point cloud classification.
In the MX9 dataset below, points are missing in the middle of the point cloud due to the trajectory of the mobile mapping scanner, as well as buildings, facades, and vertical walls blocking point collection:
When I automatically classify this dataset with the Extract Classified Point Clouds Regions command, certain parts of the point cloud are misclassified. We see that most of the road is classified as a building below. With sparse or missing point cloud data, the classification algorithm can misclassify, or not classify points at all, and could ultimately fail.
A quick workaround to this issue is to classify in chunks using the rectangle select or polygon select tool when defining your Point cloud region selection in the Extract Classified Point Cloud Regions command. Here, I will put my cursor in the Point cloud region selection box, then use the polygon select tool to highlight the right side of my road scan:
After I hit Extract, we see that the right side of the road is now correctly classified:
We can continue classifying in sections using the rectangle select or polygon select tool to obtain the best automatic classification results.
I hope this tip comes in handy next time you are working with mobile mapping data!