top of page
Orthogonal 1.png

Line Laser 3D Scanner

I designed and programmed a low cost 3D scanner that utilizes a line laser and a Raspberry pi camera to create 3D point clouds of statues and other objects. These point clouds can be used to recreate the object as a solid model.

 

I designed the device to be easy to manufacture and use. The components and frame are modular and can be 3D printed on nearly any printer. the device can be mounted on any tripod with a standard adapter and run off of standard USB power supplies.

Lidar sketch.png

Line Laser

Camera

Target

Lidar sketch 2.png

The device collects range data at many points in order to build a 3D point cloud of an object. The device has a line laser and a camera mounted a fixed distance apart. Due to the field of view angle theta of the camera, the position of a point of laser light on the target, in the camera view changes based on the distance the device is from the target. The distance of the laser point on the target can the be calculated with basic trigonometry. Since the line laser projects an infinite number of light points, the distance from the target can be determined for an entire laser line in one scan. A 3D scan of an object can be completed by changing the angle and position of the device in order to build a full point cloud. The camera takes high exposer images at each angle and position that the device is set. A script then uses computer vision to identify the location of the point of laser light in an image and calculates the range using the  method defined.  Using the angle and the relative position of the device, this range data is used to construct a 3D point cloud.

3 cups

Flour

1 cup

Sugar

2 tsp.

Oil

1½ cups

Butter

Raspberry Pi Camera

Able to detect laser with long exposure

Line Laser

Red laser

Raspberry Pi

Used for main processing

Arduino

Power distribution and auxiliary functions

Target Test Data Full.gif
3DPrint.png
Ed Turtle.png

I conducted a test of the device by scanning a statue on the UMD campus. The point cloud generated represents 3 camera locations with about 20 angles each (60 images). This raw point cloud was filtered and converted into a surface model (.stl) (in yellow). This model was able to be 3D printed as a scale model of the statue. The result was a rough proof of concept, however, with more data, filtering is more effective, and the detail increases greatly.

bottom of page