Manual classification

What is Manual Classifications

Manual classification is the task of assigning a label to each point in a point cloud. There are many applications for this, one example is building reconstruction where points need to be classified based on their z-value (height), which can then be used to create 3D models using that information. Manual Classification can also be used for any sort  of segmentation, such as separating objects from an image or to separate different types of vegetation.

The task itself is fairly straight forward and consist of placing points into a set number of categories (clusters), however deciding how many clusters there should be beforehand can be difficult as it needs to be done based on experience; those familiar with the data will  be able to make a more informed decision.

How is Manual Classifications used

This is a very flexible approach which can be used to classify data in many different ways. One example of this would be building reconstruction where you need to label each point in the point cloud with its height value. This will give you 3D models based on that information and aid in delivering it to your client in a form which is easy and  quick to make 3D prints from.

Manual Classification can also be used for any sort of segmentation, whether it is separating different types of vegetation in an image or to separate objects from each other. For example if you are taking images through a microscope, manual classification would be very beneficial because you could separate the particles into different groups depending on  colour or a specific characteristic.

How to use Point Cloud to use Manual Classifications

To use manual classification you need to load in your point cloud. You can do this using the ‘Load Point Cloud’ button on the toolbar. Then choose Manual Classification from the drop down menu, and it should run automatically. You can then begin manually classifying points by clicking on them.

Change the point that you are classifying, click on a different point and it will be highlighted. When a point is selected, a red circle will be drawn around it to show which one you have chosen.