The contact 3D scanner calculates the depth of by touching the surface of the object, such as a coordinate measuring machine, which is a typical contact 3D scanner. This method is so accurate that is often used in engineering and manufacturing industry. However, because of the scanning process must be in contact with the object, the object to be measured may be damaged by the probe. It is not suitable for high value objects such as ancient artifacts, relics and reconstruction work. In addition, compared with other methods of contact scanning taking a long time to coordinate, the latest measuring machines can complete hundreds of measurements per second. Besides, optical techniques such as laser scanners’ operation frequency is high up to ten thousand to five million times per second. Non contact active scanning refers to cast the additional energy to the object, to calculate three-dimensional spatial information by the reflection of energy. The common cast energies are general visible light, high energy beam,ultrasound and X ray.
Time delay locating method
Time delay locating method: 3D laser scanner is a kind of active scanner that uses laser light to detect the target. The lidar is a laser rangefinder with which main technique is time delay locating. The way that laser rangefinder determines the distance between the object and the surface is converse the time of the laser pulse coming and going to the measuring instrument. That is, the machine emits a laser pulse, the laser light hits the surface of the object, and then the signal is received by the detector and the time is recorded. Since the speed of light is an already known condition, the time of coming and going back can be converted into the distance that the signal travels. And the distance is twice as much as the distance from the machine to the surface of the object, so if the round-trip time for optical signal, then optical signal is equal to the distance of walking. Obviously, the 3D laser scanner of time delay locating method is of high measurement precision, which can accurately measure the time because optical signal goes 1 mm in about 3.3 picoseconds; micromicrosecond).
The laser rangefinder can only measure the distance from a single point to the machine. Therefore, if the scanner is to scan the full field of view, it is necessary to make each laser signal emitted at different angles. And this laser rangefinder can achieve this goal through its own level of rotation or internal rotation mirrors. The rotating mirror is more widely used because of its light weight, high speed and high precision. The typical time ranging type laser scanner can measure 10 thousand to 1 hundred thousand targets per second.
The triangulation ranging 3D laser scanner is a kind of active scanner which is based on laser light detecting the environment. Compared with the time delay locating method, the triangulation ranging 3D laser scanner emits a laser to the object to be measured, and uses the camera to find the laser spot on the object to be measured. With the distance of the object ranges(from the object to the 3D laser scanner), the position of the laser spot in the camera also ranges. This technique is known as the triangulation method because the laser spot, the camera, and the laser itself constitute a triangle. In this triangle, the distance between the laser and the camera, and the angle of the laser in the triangle, are known conditions. From the position of the laser spot in the camera, we can determine the angle of the camera in the triangle. These three conditions can determine a triangle, and can calculate the distance of the object to be measured. In many cases, we often replace a single laser point by a linear laser stripe for scanning which greatly accelerates the process of measurement.
The hand-held laser scanner constructs the 3D pattern by the triangulation ranging method mentioned above: with the hand-held device, it emits the laser spot or the linear laser light to the object to be measured. With two or more than two detectors (electrical coupling components or position sensing components) to measure the distance between the surface of the object and the handheld laser scanner, it usually calls for the help of a specific reference point- which is usually a sticky and reflective patch- to locate and calibrate in space. The data obtained by these scanners will be imported into the computer and then converted into 3D model by software. Hand-held laser scanner also can integrate passive scanning (visible light) data (such as the structure or color distribution of the object, ) to build a more complete 3D model of the object.
It casts one-dimensional or two-dimensional image to the object, according to the situation of the image deformation, determines the shape of the object’s surface. With which it can scan at very fast speed. Comparing with probe measuring one point at one time, this method can measure several points or a large field at one time, so it can be used in dynamic measurement.
Use a projector to project a sine wave modulated grating onto a book. Modulated light of the three-dimensional scanner continuously adjust the intensity of the light according to the time, and the commonly used modulation is a periodic sine wave. The depth of the distance can be calculated by observing the change of the brightness of each pixel and the phase difference of the light. The light source can be adjusted by a laser or a projector, and the laser light can reach a high precision. However, this method is very sensitive to the noise.
Non-contact passive scanning
The passive scanner itself does not emit any radiation (such as laser), but to measure the surrounding radiation reflected by the surface of the object to achieve the desired effect. Because of the visible radiation in the environment, it is quite easy to obtain and use. Most of these types of scanners mainly detect the visible light of the environment. But comparing with other radiation lines of visible light, such as infrared, can also be used for this purpose. Because most of the time, passive scanning method does not require very special hardware support. So these passive products tend to be quite cheap.
The traditional stereo imaging system uses two cameras placed together to look at the object to be reconstructed. This method is conceptually similar to human eyes by image perception of overlapped projections (of course human’s perception of the depth information is much more complex). If two camera focal length and distance between each other are known, and two pictures captured can be successfully coincided, then the depth information can be obtained quickly. This method must rely on the effective analysis of picture pixel matching, generally using block matching or epipolar geometry algorithm. The stereo vision method using two cameras is also called binocular vision, and the other is the extension of the three eye vision and other more cameras.
In the early stage, it was proposed by B.K.P. Horn and other scholars that the brightness value of image pixels was used to solve the pre-designed color model. The solution of equation is the depth information. Because the number of unknowns in the equations is over restrictions, it is necessary to reduce the scope of the solution set by more assumptions. For example, we can obtain the exact solution by adding the surface differentiable property, the curvature limit, the smoothness degree and the more restrictions. This method is derived from the Woodham stereo optical method.
Stereo optical method
In order to compensate for the lack of information provided by photometric method of a single photograph, stereo optical method using one camera shooting multiple photos, the photo shoot angle is the same. The difference is the lighting conditions. The simplest stereo optical method uses three light sources to illuminate the object from three different directions, and each time only opens one light. After the completion of the shooting, then combine with three photos and use perfect diffusion model of optical to solve the gradient vector of the surface of the object. After the integration of vector field, three-dimensional model can be obtained . This method is not applicable to the smooth and approximate surfaces similar to Lambertian.
This method uses a series of objects to form a three-dimensional contour lines. When some of the surface of the object can not be displayed online, the 3D information will be lost. The common way is to place the object on a rotating electric turntable. Shoot the image in each small angle turning, and then remove the background through image processing techniques and collect the contour lines of every angle. Finally, the 3D models are “draw”.
In addition, there are some methods requiring the user to provide information in the reconstruction process, with the help of the unique performance of the human visual system to help reconstruct program. These methods are based on the principle of photo photography that it shoots images for the same object to calculate three-dimensional information. Another similar way is panoramic reconstruction, which is to take surrounding pictures on a fixed point.