Table 2 A comparison of 3D imaging techniques

Active methods

3D Laser Scanner (LiDAR)

Determines the range by targeting an object or a surface using a laser to measure the time for the reflected light to return to the receiver

(1) High resolution

(2) Low-cost devices available

(3) Relatively insensitive to lighting conditions

(4) Suitable for high volume scanning

(1) Complex scanning process

(2) Requires calibration

(3) Lack of depth detection mechanism

(1) Fields

(2) Controlled environments

Laser Triangulation

Captures 3D data by pairing the laser light source with a sensor array (e.g., camera)

(1) High accuracy at close range

(2) High resolution

(3) Low cost

(4) Relatively insensitive to the lighting conditions or surface textures

(1) Low accuracy at large distances

(2) Computationally intensive

(3) No color information

(4) Difficult to scan transparent or reflective surfaces

(1) Fields

(2) Controlled environments

Photometric Stereo (PS)

Estimates local surface orientation by using a sequence of images of the same surface from the same viewpoint but under illumination from different directions

(1) High resolution

(2) Low cost

(3) High speed

(4) Reasonable computational cost

(1) Difficult to handle shiny and semi-translucent surfaces

(2) Requires calibration

(1) Fields

(2) Controlled environments

Structured Light

Projects a series of known patterns onto the object and builds up a 3D image by measuring the deformation of the patterns

(1) High accuracy

(2) High-speed scanning

(3) High resolution

(1) Vulnerable to ambient light interference

(2) Decreasing accuracy as the measurement distance increases

(1) Controlled environments

Time of Flight (ToF)

Builds up a 3D image using light emitted by a laser or LED source and measuring the roundtrip time between the emission of a light pulse and the reflection

(1) Easy setup due to the small size of the camera

(2) High-speed measurement

(3) Wide measurement range

(4) Relatively insensitive to ambient light

(1) Expensive

(2) Difficulties with shiny surfaces

(1) Fields

(2) Controlled environments

Tomographic Methods

Create a series of 2D slices in order to generate a 3D volume

(1) High resolution

(2) High accuracy

(1) Not suitable for large-scale field studies

(2) Time-consuming

(3) High cost

(1) Controlled environments

Passive methods

Light Field Measuring

Reconstructing the 3D information of a scene by capturing both the radiant intensity and the direction of the incoming light in one plane

(1) High-dimensional representation

(1) Technical complexity

(2) Computational cost

(1) Fields

(2) Controlled environments

Multi-view Stereo (MVS) Techniques

Uses two or more cameras to generate parallax from different perspectives to obtain distance information about an object

(1) Simplicity of use

(2) Low cost

(3) High accuracy

(1) Limited imaging range

(2) Low imaging quality

(3) Computationally intensive

(4) Poor performance in real-time

(5) Relatively sensitive to the environment conditions

(1) Fields

(2) Controlled environments

Space Carving

Uses a voxel grid and information from pictures taken from different perspectives to remove voxels which are not part of the object

(1)Avoids correspondence problem

(2) Low cost

(2) Requires fewer images than, e.g., SfM to obtain a good representation

(1) Calibration requirements

(2) Quality of reconstruction depends on the number of views

(3) Exact segmentation requirements

(4) Lack of depth detection mechanism

(5) Not suitable for highly non-convex objects

(1) Controlled environments

Structure from Motion (SfM)

Captures 3D information from sequences of overlapping 2D images

(1) Low cost

(2) High resolution

(3) High color reproduction

(4) Wide measurement range

(5) Mature algorithm, frequently used

(6) Some degree of automatic calibration

(1) Need to move measurement equipment continuously

(2) Reconstruction complexity

(3) Time-consuming

(1) Fields

(2) Controlled environments