Structure Capture App Settings Guide
Main Screen Settings
Setting Name | Setting Function | General Recommendation |
Mesh Detail | This sets the Depth map voxel density/size, which increases/decreases mesh detail at the cost of processing power. This affects image resolution and noise in your scan. (Voxels are little cubes that fill up the bounding box. If you have too many voxels, then your iPad will slow down, heat up and become laggy, so more is not always better) CAUTION: Any value below 1 mm can quickly eat all available memory |
For smaller subjects, we recommend using a voxel size between 1-3mm for optimal results. Smaller voxels capture more details. (Note that smaller voxel sizes require more memory, more computational power (the scanning becomes laggy as the CPU can't keep up), and the resulting scan is noisier. Also, the resolution of the final mesh is usually better. For larger subjects, we recommend increasing the voxel size >3mm to the point in which you are happy with the results. Larger voxel sizes average out noise more effectively. |
Color+Depth vs Depth only | This specifies the tracking algorithm and determines the system the sensor and software use in order to maintain its position in space. Color + Depth utilizes both the Structure Sensor camera and iPad camera in order to generate the mesh scan. Depth only utilizes only the Structure Sensor camera. | We generally recommend enabling Color + Depth unless you know the scene has limited features (different colors, etc) to track against. In that case, then we recommend Depth only |
File Format | This allows you to select which file format you want your scan to be exported to. OBJ files are exported as Zip files, which includes 3 files: the mesh, material, and texture. PLY and STL files are exported as just the mesh. | We generally recommend selecting OBJ files as most post-processing softwares are able to import these files, but it ultimately depends on which file format you require. OBJ files are commonly used in 3D modeling applications and are supported by many software tools. They are versatile and suitable for transferring complex 3D models. (Note: this is the only format that supports texture). PLY files are used in a variety of applications, including computer graphics, computer-aided design (CAD), and 3D scanning. They are particularly useful for storing point cloud data. (Note: this file type will not show texture/color). STL files are commonly used in 3D printing and rapid prototyping. They are also used for transferring surface geometry between different 3D modeling software. (Note: This file type will not show texture/color). |
Bounding box | This sets the scene/area for tracking and scanning. Use ONE finger to slide the bounding box to the center of the object, toward and away from you. Then use TWO fingers to resize the bounding box to make sure it is larger than the object. |
We recommending setting the bounding box to be slightly larger than the object you are trying to scan. The larger the bounding box, the lower the resolution of the final scan. |
Gear Icon Settings
Setting Name | Setting Function | General Recommendation |
Record OCC | Turning this on creates OCC files during scanning, which can then be sent to support for analysis. OCC files are special diagnostic files that include mesh, depth maps, and IMU data. | We recommend leaving this as the default "OFF" in most circumstances as these files can be quite large, which could affect processing power. |
Sensor | This determines whether scanning is done using a Structure Sensor or the iPad/iPhone's TruDepth camera. | We recommend scanning with a Structure Sensor for best performance and accuracy. |
High resolution color | Turning this "ON" selects the highest available resolution for color frames (1280x720). Turning this "OFF" selects VGA resolution for color frames. Turn off if color meshes are not needed. |
Turn "ON" High resolution color, but in situations where color is irrelevant, turn this feature off to save processing power. Overall, it's usually safe and recommended to use High resolution color as it should look better on the screen and results in better texture in most cases. |
Lock autofocus | Turning this "ON" locks focus after scanning has started. Turning this "OFF" keeps autofocus on during the scanning process. | Turn "ON" Lock autofocus |
Preset (See Below) | (See Below) | (See Below) |
Default Confidence Threshold | This sets Depth Confidence, a tool used to remove pixels from a depth frame likely to be inaccurate. Higher values remove more pixels, but reduce tracking. WARNING: Any value above 7 can drastically reduce tracking. | In most cases, turn "ON" Default Confidence Threshold. However, if you experience scan quality issues, then move this down between 6 to 3 to start. |
Gain | This specifies IR frame digital gain and gives you more finite control over infrared gain (sensitivity to infrared). Gain settings selected here will remain constant throughout the scan. In situations of low ambient infrared, increasing gain can improve capture. Higher gain is preferred in situations of low ambient infrared. Lower gain is preferred in situations of high ambient infrared. | In most cases, this does not need to be adjusted. Otherwise, this setting might be relevant in case you need to scan something far away from the sensor. However, the sensor presets already take this into account, so make sure you do not override the sensor preset with the gain settings. |
Position | This determines how the bounding box positions itself. 'User Defined' lets the user set the bounding box size and distance from the sensor. 'Auto (Table Top)' automatically detects a ground plane and positions the bounding box on top of it (only the bounding box size is editable). |
In most cases, it is best to select "User Defined" to allow you to set the distance and size of the bounding box. Otherwise, you can use 'Auto (Table Top)' if a ground plane is easily trackable. like on a table top. |
Align with Camera | Turning this "ON" will make it easier to scan things floating in the air. For example, a foot floating in the air will be easier to scan with this turned on. Turning this "OFF" searches for, adjusts, and aligns Structure Sensor to the ground plane/ground surface. (Note that turning this off and scanning an object in the air will be more difficult because the bounding box might shift to the bottom of something other than what you are trying to scan.) |
In most cases, turn "ON" Align with Camera |
Mode | This determines the tracking mode for your sensor. 'Object' is ideal for subjects that are ~6 foot / 2 meters or smaller 'Turntable' is good if the subject is on a lazy susan or similar and the sensor is in a fixed position 'Area' is optimal if you are doing first person point of view scanning |
In most cases, this does not need to be adjusted from 'Object." |
Type | This specifies the tracking algorithm and determines the system the sensor and software use in order to maintain its position in space. Color + Depth utilizes both the Structure Sensor camera and iPad camera in order to generate the mesh scan. Depth only utilizes only the Structure Sensor camera. | We generally recommend enabling Color + Depth unless you know the scene has limited features (different colors, etc) to track against. In that case, then we recommend Depth only. |
Mesh Detail | This sets the Depth map voxel density/size, which increases/decreases mesh detail at the cost of processing power. This affects image resolution and noise in your scan. (Voxels are little cubes that fill up the bounding box. If you have too many voxels, then your iPad will slow down, heat up and become laggy, so more is not always better) CAUTION: Any value below 1 mm can quickly eat all available memory |
For smaller subjects, we recommend using a voxel size between 1-3mm for optimal results. Smaller voxels capture more details. (Note that smaller voxel sizes require more memory, more computational power (the scanning becomes laggy as the CPU can't keep up), and the resulting scan is noisier. Also, the resolution of the final mesh is usually better. For larger subjects, we recommend increasing the voxel size >3mm to the point in which you are happy with the results. Larger voxel sizes average out noise more effectively. |
File format | This allows you to select which file format you want your scan to be exported to. OBJ files are exported as Zip files, which includes 3 files: the mesh, material, and texture. PLY and STL files are exported as just the mesh. | We generally recommend selecting OBJ files as most post-processing softwares are able to import these files, but it ultimately depends on which file format you require. OBJ files are commonly used in 3D modeling applications and are supported by many software tools. They are versatile and suitable for transferring complex 3D models. (Note: this is the only format that supports texture). PLY files are used in a variety of applications, including computer graphics, computer-aided design (CAD), and 3D scanning. They are particularly useful for storing point cloud data. (Note: this file type will not show texture/ color). STL files are commonly used in 3D printing and rapid prototyping. They are also used for transferring surface geometry between different 3D modeling software. (Note: This file type will not show texture/color). |
Presets
Preset Name | Preset Function | General Recommendation |
Body | For scanning body parts/limbs like a foot Ideal distance (m) - Min: 0.36 / Max: 1.02 |
Choose this preset if you are intending to scan an arm, a foot, or any other object that requires a free-floating, groundless bounding box Smallest scan size: Hand/wrist Ideal scan size: Lower limb Largest scan size: Torso |
Generic | For general scanning; this is a good middle ground for indoor scanning. The largest limitation is that it needs to be used indoors, and its minimum focal length is slightly longer than the Body Preset. Ideal distance (m) - Min: 0.453 / Max: 1.005 |
Choose this preset most of the time Smallest scan size: Foot Ideal scan size: Torso Largest scan size: Body |
Outdoor | For scanning objects where there is direct or indirect sunlight on the scene either outdoors or indoors if the object is large (e.g. a head or larger) Ideal distance (m) - Min: 0.56 / Max: 12+ |
Choose this preset when scanning outdoors or where direct/indirect sunlight is on the scene Smallest scan size: Mailbox Ideal scan size: ATV Largest scan size: Tractor |
Room | For scanning objects larger than a person or a large area Ideal distance (m) - Min: 0.56 / Max: 12+ |
Choose this preset when scanning an entire person's body or a large object/area Smallest scan size: Doorframe Ideal scan size: Wall section (e.g. fireplace) Largest scan size: Small room (e.g. office, kitchen, bedroom) |
Close Range | For scanning objects smaller than a foot at close range Ideal distance (m) - Min: 0.325 / Max: 0.75 |
Choose this preset for small objects at close range Smallest scan size: Hand/wrist Ideal scan size: Grapefruit Largest scan size: Head |
Hybrid | For scanning both close range and long range, but it may be less accurate than either the Default preset or the Body preset. This preset increases range at the expense of depth resolution. Ideal distance (m) - Min: 0.355 / Max: 5.89 |
This is an ST02-specific preset and is selected based on what distance range you intend to scan from Smallest scan size: Torso Ideal scan size: Body (at distance of 2+ meters) Largest scan size: Book case |
Dark Object | For scanning dark objects and/or if indoors and the object is small (e.g. smaller than a head). This preset allows for significantly better detection of dark and reflective objects Ideal distance (m) - Min: 0.67 / Max: 5.72 |
Choose this preset when scanning a dark or reflective object Smallest scan size: Dark Grey mug Ideal scan size: Dark Grey teapot Largest scan size: Dark Grey chair |
Medium Range | For scanning for larger objects or smaller areas Ideal distance (m) - Min: 0.56 / Max: 10+ |
This preset is selected based on what distance range you intend to scan from Smallest scan size: Torso/lower body Ideal scan size: Body Largest scan size: 2-person couch |
Recommended Settings for Specific Use Cases:
For scanning smaller objects (e.g. a foot or head) we recommend the following combination of settings:
Setting Name | Setting Status |
Record OCC | OFF |
Sensor | Structure Sensor |
High resolution color | ON |
Lock autofocus | ON |
Preset | Body |
Default Confidence Threshold | ON |
Gain | 4x |
Position | User Defined |
Align With Camera | ON |
Mode | Object |
Type | Color+Depth |
Mesh Detail | 2.5mm |
File Format | Obj (Zip) |
For scanning medium-sized objects (e.g. a backpack), we recommend the following combination of settings:
Setting Name | Setting Status |
Record OCC | OFF |
Sensor | Structure Sensor |
High resolution color | ON |
Lock autofocus | ON |
Preset | Generic |
Default Confidence Threshold | OFF; set to 3 |
Gain | 4x |
Position | User Defined |
Align With Camera | ON |
Mode | Object |
Type | Color+Depth |
Mesh Detail | 3.0mm |
File Format | Obj (Zip) |
For scanning larger objects (e.g. an entire person's body), we recommend the following combination of settings:
Setting Name | Setting Status |
Record OCC | OFF |
Sensor | Structure Sensor |
High resolution color | ON |
Lock autofocus | ON |
Preset | Generic |
Default Confidence Threshold | OFF; set to 6 |
Gain | 4x |
Position | User Defined |
Align With Camera | ON |
Mode | Object |
Type | Color+Depth |
Mesh Detail | 3.0-5.0mm |
File Format | Obj (Zip) |