3D Scanner & Post Processing

This assignment is about asking us to compose a model by using scanners.

This week we learned how to use 3D scanning to directly transform a concrete object into a processing-able model with the assistance of 2 type of technologies, namely photogrammetry and Einscan handheld 3D scanner.

However, there's suddenly something wrong with the handheld 3D scanner, as a result, we turned to use the photogrammetry, which is a photo processing application that is capable to generate 3D models with multiple photos taken from diverse angles around the aim objects.

It's worth noting that we cannot move the object once we decide where to set the object, since photogrammetry will try to detect in terms of the the surrounding relative position. Also, it's better to set the object in a place with some obvious reference substances, rather than in a plain background.

Though Professor reminds us that smartphones are not best choice when taking photos, we still firstly try to take photos with phones. Without expectation, photogrammetry denied the quantity of those photos because of their relatively resolution ratio. As we could tell the appearance of a red warning sign at the upper left corner of those uploaded photos, photogrammetry would intend for higher-quality photos to process.

Also, the number of photos I took was fewer than it needs.

I went to IT for a SLR camera to take those photos. As you can tell from the upper left conners this time, the sign turned to be green, which means their quality is satisfactory.

However, the other red signs at the right upper conner means that photogrammetry cannot properly use it. Later I figured out some of them might be out of focus, or be relatively lacking in lightness.

Then I made the third trail with over 100 photos. This time the outcome is relatively satisfying. However, since photogrammetry composes models with dots, I thought I could achieve a better version so I uploaded more detailed photos to try.

Since there were about 150 photos to be in process, it took about an hour and a half for photogrammetry to construct the model.

The final effect of this model is not so bad, since the whole model was composed of dots.

However, I believe the final model would be better if the scanned model is a bit larger.

It is so important to select a place to set the model, since it would be hard to compose the model if the background place is too bright or too dim. Professor Andy recommended that it would be a good choice if we could take the pictures on a cloudy day outdoors.

Also, it's worth while moving step by step to take photos from every angle around the model as the photogrammetry desires huge amount of details from photos to generate the model.

Finally, it would be better if we could set the model on a relatively complicated background, so that photogrammetry could use some of those background information as a reference.