Taking the photographs

3DSOM Pro requires photographs taken from around 30 angles, to create a 3D model of an object. (You may also need to take additional photos to calibrate your camera lens — see the Calibrating the Lens tutorial.)

General points to note...

TIP:

Frame the view so that the calibration target(s) takes up about 1/3 of the total image.

Shooting with the calibration target(s) in every frame

If using a computer controlled turntable you will typically take sequences of 16 shots for each camera with the object rotating between each shot.

If you are manually rotating the object (e.g. on a lazy susan) you can take a sequence of 15 shots by lining up each row of dots with a reference point.

The camera should be looking slightly downwards to ensure the calibration target(s) is clearly visible as shown in Figure 1. If you are using multiple cameras then the lower camera should be at an elevation angle of around 15 to 20 degrees relative to the object and the next camera should by 15 to 20 degrees above that (30 to 40 degree elevation to the object). You can also add a 3rd camera at a higher elevation angle of around 65 degrees. The projector should be mounted between the lower two cameras or for 3 camera systems close to the middle camera.

Figure 1. Correct Camera position for the lower camera angle

The photographs should be taken from different viewpoints by rotating the mat either clockwise or anti-clockwise into 15 or 16 different positions.

Make sure that the bottom of your backdrop cuts through the 3DSOM Pro stand and not the object (see Figure 2).

Figure 2. Correct height to frame photographs from the side

For best results, you will need to take a number of sequences (complete rotations)...

A basic 3D model can be created without using a stripes scan just using the silhouette information. This can be useful for modeling difficult objects such as semi-transparent glass or mirrored objects.

TIP:

If you only have one camera then you will need to take the stripe shots with the camera at the lower position (15 degree elevation). Then carefully raise the camera to the higher position (30 degree elevation) and take a second sequence from exactly the same rotation angles. If you don't have a computer controlled turntable, you can move the camera for each rotation position - e.g. take the 1st low angle shot, move the camera to take the 1st higher angle shot, then rotate the object and mat and take the higher and then lower shots and repeat.

To capture the top of the object you can take a single high-up shot to capture the appearance of the object from above - make sure you can still see at least 4 complete rows of dots on the calibration target(s).

Once you have captured all the images you can take a single shot of the bottom by moving the object (see Adding texture underneath an object). Alternatively take a second complete set of images with the object on it's side or upside-down and merge the two scans (see Merging multiple scans).

Shooting with separate calibration shots

If you are using our partner hardware systems with computer controlled turntable you can eliminate the need for keeping the calibration target(s) in every shot. This allows you to create more automated masks and you don't need to keep any residual lighting turned on when capturing the stripe shots.

The bundled control software will ensure that the calibration shots, normal shots and stripe shots are suitably named so that 3DSOM Pro can process them correctly. See the example projects to see the image filenaming convention used. Each camera position should have a unique name with a matching calibration shot with suffix "_cal.jpg".

For a hardware system with auto-PNG masking and/or a transparent platform, you also won't need to use the stand.

Workflow...

Saving the photographs

Once all the photographs have been taken, they should be saved in a single folder for use with 3DSOM Pro.

3DSOM Pro accepts images in JPEG, PNG, PPM, BMP or TIFF formats.