INDEX

1) Abstract 2) Introduction..... 3) Purpose 4) Scope 5) Overall Description 5.1) Concept.. 5.2) Principle. 5.3) Components Of TOF camera 5.4) Working 5.5) Faults In TOF camera. 5.6) Practical Use 5.7) Technologies Developed 5.8) Advantages 5.9) Short Comings 5.10) Applications 6) Conclusion 7) Bibliography

ABSTRACT

A traditional camera delivers a two-dimensional image of the surrounding world. The 3D-time-of-flight camera (ToF) - which has been developed recently captures not only two, but three dimensions in one step by measuring the distance from each pixel of the camera chip to the object. This way a large number of coordinates is determined for each exposure a so called frame. More than 20 frames per second may be taken. The result is a point cloud similar to the familiar data gathered via terrestrial laser scanning (TLS). However, some fundamental differences concerning gathering data and interacting with the instrument have to be discussed. Since handling is as simple as using a 2D-camera - making new ways of 3D-recording possible - there is huge potential for the development of new userfriendly measuring tools. Especially the introduction of the 3D-camera as a surveying instrument for architectural documentation is of interest. However, the particular method utilized to generate the point-cloud raises seriousquestions regarding the quality of the data. All information is presented here with the further development of the 3D-camera in mind. The on-hand experience gained employing the current generation 3D-cameras can be applied to subsequent, improved generations. A major intention of this paper is to familiarise the reader with the
2

new device, its problems and possibilities for its practical application.

INTRODUCTION

What is Time Of Flight?

Time of flight (TOF) describes a variety of methods that measure the time that it takes for an object, particle or acoustic, electromagnetic or other wave to travel a distance through a medium. This measurement can be used for a time standard (such as an atomic fountain), as a way to measure velocity or path length through a given medium, or as a way to learn about the particle or medium (such as composition or flow rate). The traveling object may be detected directly (e.g., ion detector in mass spectrometry) or indirectly (e.g., light scattered from an object in laser doppler velocimetry). Time-of-flight cameras are relatively new devices, as the semiconductor processes have only recently become fast enough for such devices. The systems cover ranges of a few meters up to about 60 m. The distance resolution is about 1 cm. The lateral resolution of time-of-flight cameras is generally low compared to standard video cameras, at 320 240 pixels or less. Only one camera reports 484 x 648 pixels of resolution using a standard CCD sensor. The biggest advantage of the cameras may be that they provide up to 100 images per second.

Why to use a 3-D TOF camera?

When using a 3D-camera three specific reference values are of increased interest: the accuracy of coordinates, the resolutionof the recording chip (number of pixels) the maximumrange procurable. With the development of the latest generation of 3D-cameras, much progress has been made in regard to thesethree items: accuracy is now in the range of a few centimetres, resolution goes up to 40k pixels the maximum distance wasraised to about 20m.

In light of these developments the question , whether the 3D-camera may be used as an instrument for surveying and recording is worth investigating. This may be done by extrapolating from the technical specifications of currently available 3Dcameras into the near future, which will most certainly bring higher resolution, increased accuracy and an expanded maximum distance. Further developments of this kind will be realised soon as prospective customers exist in the field of robotics, vehicle construction and many more areas. The main questions are as follows: 1. Are there any fundamental problems in the process of recording architecture? 2. Which new possibilities for recording of geometry and for visualisation purposes arise for practical work?
4

3. Is the proposed 2D- / 3D-system capable to replace current devices?

PURPOSE

Today, is world of technology. None of the technologies available in any field at present is complete in itself. The scientists are continuously trying to develop new one which would be better than the present one.

The 2-D camera has been used for developing pictures, video recording, graphics modulation, and in many applications for many decades, but now a 3-D time of flight technology has been developed. A traditional camera delivers a two-

dimensional image of the surrounding world. The 3D-time-of-flight camera (ToF) which was developed in the last decade - captures not only two, but three dimensions in one step by measuring the distance from each pixel of the camera chip to the object. This way a large number of coordinates is determined for each exposure a so called frame.

The 3-D time of flight technology is quiet sufficient in producing efficient image than a 2-D technology. The target source is illuminated using infra red light ( generally), the reflected light is gathered by the image sensor and a cloud of pixels is generated, each point corresponds to every exposed part of the object.

The main advantage of using this technology is the speed it provides. Time-offlight cameras are relatively new devices, as the semiconductor processes have only recently become fast enough for such devices. The systems cover ranges of a few meters up to about 60 m. The distance resolution is about 1 cm. The lateral resolution of time-of-flight cameras is generally low compared to standard video cameras, at 320 240 pixels or less. Only one camera reports 484 x 648 pixels of resolution using a standard CCD sensor. The biggest advantage of the cameras may be that they provide up to 100 images per second. Another main feature of TOF camera is that it is quiet simple to use them. As the whole system is very compact, the illumination is placed just next to the lens, and no mechanical moving parts are needed. The 3-D TOF camera provides a better algorithm to calculate distance between object and the light source. Therefore, this task uses only a small amount of processing power, again in contrast to stereo vision, where complex correlation algorithms have to be implemented. After the distance data has been extracted, object detection, for example, is also easy to carry out because the algorithms are not disturbed by patterns on the object. So, 3-D time of flight technology has a great advantage over the traditional 2-D technology in image generation and it will be quiet useful in upcoming days.

SCOPE

Due to fast speed and simplicity of 3-D time of flight technology, it has become very popular in many fields. Various technologies using 3-D TOF technology: Automotive applications Time-of-flight cameras are also used in assistance and safety functions for advanced automotive applications such as active pedestrian safety, precrash detection and indoor applications like out-of-position (OOP) detection. Human-machine interfaces / gaming As time-of-flight cameras provide distance images in real time, it is easy to track movements of humans. This allows new interactions with consumer devices such as televisions. Another topic is to use this type of cameras to interact with games on video game consoles. Measurement / machine vision

Other applications are measurement tasks, e.g. for the fill height in silos. In industrial machine vision, the time-of-flight camera helps to classify objects and help robots find the items, for instance on a conveyor. Door controls can distinguish easily between animals and humans reaching the door.

Robotics Another use of these cameras is the field of robotics: Mobile robots can build up a map of their surroundings very quickly, enabling them to avoid obstacles or follow a leading person. As the distance calculation is simple, only little computational power is used.

OVERALL DESCRIPTION

1)Concept of Time Of Flight

Time-of-Flight (ToF) sensors provide a direct way for acquiring 3D surface information of objects. More recently, applications like gesture recognition or automotive passenger classification are using ToF sensors and ToF is on its way to become a component of consumer electronics. As ToF sensors provide data at rates higher than 15 Hz, they are suitable for real-time 3D imaging and can also be used as an additional imaging modality in medicine. We are convinced that ToF technology can contribute to enhance applications within medicine and suggest several applications within this field. Time-of-flight cameras are relatively new devices, as the semiconductor processes have only recently become fast enough for such devices. The systems cover ranges of a few meters up to about 60 m. The distance resolution is about 1 cm. The lateral resolution of time-of-flight cameras is generally low compared to standard video cameras, at 320 240 pixels or less. Only one camera reports 484 x 648 pixels of resolution using a standard CCD sensor. The biggest advantage of the cameras may be that they provide up to 100 images per second.

2) Time Of Flight Principal

A Time of flight camera (TOF camera) is a camera system that creates distance data with help of the following principle:

Time of flight Principle (TOF) describes a variety of methods that measure the time that it takes for an object, particle or acoustic, electromagnetic or other wave to travel a distance through a medium. This measurement can be used for a time standard, as a way to measure velocity or path length through a given medium, or as a way to learn about the particle or medium. The traveling object may be detected directly or indirectly.

The 3D Time-Of-Flight (ToF) camera, simultaneously delivers gray-level

images and 3D information of the scene. Illumination source is in the form of matrix of LEDs. On the basis of reflected ray: Distance Image (D(i, j)) is computed based on the phase shift between the emitted and reflected signals, Amplitude Image (A(i, j)) is estimated based on the amplitude of the reflected signal at every pixel location.

10

3) Components Of Time Of Flight camera

A time-of-flight camera consists of the following components:

Illumination unit: It illuminates the scene. As the light has to be modulated with high speeds up to 100 MHz, only LEDs or laser diodes are feasible. The illumination normally uses infrared light to make the illumination unobtrusive.

Optics: A lens gathers the reflected light and images the environment onto the image sensor. An optical band pass filter only passes the light with the same wavelength as the illumination unit. This helps suppress background light.

Image sensor: This is the heart of the TOF camera. Each pixel measures the time the light has taken to travel from the illumination unit to the object and back. Several different approaches are used for timing; see types of devices above.

Driver electronics: Both the illumination unit and the image sensor have to be controlled by high speed signals. These signals have to be very accurate to obtain a high resolution. For example, if the signals between the illumination unit and the sensor shift by only 10 picoseconds, the distance changes by 1.5 mm. For comparison: current CPUs reach frequencies of up to 3 GHz, corresponding to clock cycles of about 300 ps - the corresponding 'resolution' is only 45 mm.

Computation/Interface: The distance is calculated directly in the camera. To obtain good performance, some calibration data is also used. The camera then provides a distance image over a USB or Ethernet interface.
11

4) Working of TOF camera

The simplest version of a time-of-flight camera uses light pulses. The illumination is switched on for a very short time; the resulting light pulse illuminates the scene and is reflected by the objects. The camera lens gathers the reflected light and images it onto the sensor plane. Depending on the distance, the incoming light experiences a delay. As light has a speed of c = 300,000,000 meters per second, this delay is very short: an object 2.5 m away will delay the light by:

The pulse width of the illumination determines the maximum range the camera can handle. With a pulse width of e.g. 50 ns, the range is limited to

These short times show that the illumination unit is a critical part of the system. Only with some special LEDs or lasers is it possible to generate such short pulses. The single pixel consists of a photo sensitive element (e.g. a photo diode). It converts the incoming light into a current. In analog timing imagers, connected to the photo diode are fast switches, which direct the current to one of two (or several) memory elements (e.g. a capacitor) that act as summation elements. In digital timing imagers, a time counter, running at several gigahertz, is connected to each photo detector pixel and stops counting when light is sensed.
12

In the diagram of an analog timer (given below), the pixel uses two switches (G1 and G2) and two memory elements (S1 and S2). The switches are controlled by a pulse with the same length as the light pulse, where the control signal of switch G2 is delayed by exactly the pulse width. Depending on the delay, only part of the light pulse is sampled through G1 in S1, the other part is stored in S2. Depending on the distance, the ratio between S1 and S2 changes as depicted in the drawing. Because only small amounts of light hit the sensor within 50 ns, not only one but several thousands pulses are sent out (repetition rate tR) and gathered,

13

thus increasing the signal to noise ratio. After the exposure, the pixel is read out and the following stages measure the signals S1 and S2. As the length of the light pulse is defined, the distance can be calculated with Medina's formula:

In the example, the signals have the following values: S1 = 0.66 und S2 = 0.33. The distance is therefore:

In the presence of background light, the memory elements receive an additional part of the signal. This would disturb the distance measurement. To eliminate the background part of the signal, the whole measurement can be performed a second time with the illumination switched off. If the objects are further away than the distance range, the result is also wrong. Here, a second measurement with the control signals delayed by an additional pulse width helps to suppress such objects. Other systems work with a sinusoidally modulated light source instead of the pulse source. Different denotations exist as far as 3D-cameras are concerned: The TOF-camera, The PMD-camera The range imaging camera (RIM or flash radar camera)

14

They are all targeted at the same type of camera, which is capable of not only detecting intensity like traditional analogue or modern digital cameras, but also the distance of each pixel to the mapped part of the object. A more precise description for the term range imaging camera is the term ToF (TOF)-camera indicating that an - optical - signal that has been emitted from the camera is reflected from the object, and also indicating that the reflected part of the emitted signal is then returned to the camera.

The flight-time of the signal on its way to the object and back is measured. The ToF is proportional to the distance. However, this is only a very general explanation of electronic distance measurement (EDM). Often a modified method (phase modulation) is used that relies on the modulation of the signal strength (the amplitude of the emitted light).

15

The wavelength of one modulation period depends on the modulation frequency f. It is calculated with the actual - velocity of light c according to Wavelength = C/F. Most 3D-cameras emit IR-light modulated with f = 20 MHz. In the example, the corresponding wavelength of 15m (15m = 300000km/s / 20 x 106/s) is reflected at a distance of about 22m. As the light passes to the object and back it runs the double distance, so the result needs to be divided by two.

The so-called scale of the distance meter is 15m/2 = 7,5m. The whole distance is d = (n x wavelength + phase), where n is the even multiple of wavelength , and phase is the rest of the wavelength which has to be measured, the so called phase difference between the emitted and the returned signal. Thus it follows that when a single modulation frequency is utilized, a direct distinct measurement is possible only up to 7.5m. So a distance of 8.05m would show as 0.55m.

This is a general description how distances can be measured by means of an amplitude modulated signal. The phase measurement the determination of . - may be done in different ways. In a 3D-camera it is done using the PMD-effect (Photonic-Mixer-Device (Schwarte)). In this case phase is represented by the difference in charge of two photo-sensitive areas of each respective pixel. Daylight may cause an overload of one of these areas.

This may cause serious errors in the distance measured by a ToF-PMD camera, making daylight suppression a fundamental requirement for any 3Dcamera to be used for surveying.

16

Above figure shows the difference of images obtained from 2-D & 3-D camera. It can be shown clearly that images obtained from 3-D camera are of high resolution than 2-D camera. 2-D camera produces redundant data which makes the image blurred and the image produced is of low resolution.

Some General Remark

In order to answer the aforementioned main questions it has not been necessary to utilize the very latest generation of 3Dcameras. However, the device used for testing needed to meet some minimum requirements, which are sufficient short-term temperature-stability as well as adequate daylight suppression. Due to its high performance in the latter area a 3k camera of PMD-technologies ( Siegen, Germany) was selected. This model, however, is soon to be succeeded by the camera 2.0 (41k). Figure 1 shows 2.0 - currently being the camera with the highest pixel-resolution available - 205 x 205 pixels - as well as the model PMD 3k with a 2D-colour-camera placed on top.

17

Figure a: 3D-camera 2.0, 41k (PMD-Technologies) Figure b: 3D-camera 3k (PMD-Technologies) mounted with 2D-camera

On both sides of the 3D-camera's cubes diode arrays are mounted emitting an amplitude-modulated IR-light that illuminates the whole scene to be captured. Apart from the distances (resp. coordinates) and the grey-values (intensities) the signal-amplitudes of the returned signals are registered for each pixel as well. Additionally, the 2D- / 3D-camera system of Bochum University delivers a colourimage with higher resolution. The lateral positions of the pixels on the chip of the 3D-camera and of the colour-camera respectively present polar coordinates together with the distances. They allow the calculation of orthogonal coordinates of the object.

18

5) Typical Faults Of TOF Camera

The emitted light illuminates the whole 3D-scene at once. This is a very important difference to common electronic distance measurement (EDM). At EDM the distance measuring ray is precisely focused. The emitted signal amplitude is D (t0) = Acos wt(0). The remitted light signal comprises the distance information in terms of a phase delay of the emitted signal: d(t) = k + a cos (wt + phase). The phase delay contains the distance information. The remitted signal amplitude a is generally much smaller than the sent amplitude A. Additionally, the returning signal may be superimposed with a term k which contains various very different kinds of influences on the signal, like, for instance, a delay that might be caused by daylight or by temperature effects.

In order to understand the difficulties concerning the application of 3Dcameras for architectural recording it is necessary to take a closer look at the instruments faults. Most faults are well-known either from the field of photogrammetry e.g. distortion of pictures or from electronic distance measurement (EDM) - like phase shifts caused by optical or electrical superimposition of the direct signal with error signals (crosstalk). Most of the faults of 3DPMD- cameras can be sufficiently modelled in a calibration process.

Certain errors may also be suppressed, e.g. daylight via the SBIcircuit (Suppression of Background Light Intensity). The short time random inaccuracy of neighboring pixels may nowadays be estimated to generally not exceeding 1cm in a distance-range of some meters. However, two fundamental influences on the
19

distances have a systematic nature. They can considerably compromise the general accuracy. Both are resulting from unwelcome reflections of the distance measuring signals. The first influence may be characterized as internal superimposition, the second as external superimposition. Superimposition in general terms means that the periodic amplitude modulated signal, which is characterized by frequency, amplitude and phase , is superimposed by a signal of the same frequency, but with another phase and generally much lower amplitude.

By this process the resulting signal may deviate significantly from the undisturbed one in regards to phase. The amount of the phase shift depends on the phase difference between both signals as well as on the ratio of the amplitudes of the signals as in the following figure:

Signal a with phase p is superimposed by a small reflected- signal b with phase p'. Resulting is a signal with phase p + t. The influence of b depends on its individual phase. A phase shift is equivalent to an error in distance. The grave influence of internal and external superimposition respectively on distances
20

measured with the 3D-camera has to be regarded in more detail in order to access the possibilities to use the tool for architectural recording. Internal superimposition effecting distance shadows that encircle the foreground like a small ribbon.

In literature internal superimposition is also characterised as scattering: Small parts of incoming signals are reflected in the camera itself, overlaying the directly reflected incoming signals on their way to the pixel. Thus the stronger direct signal is changed in phase by the reflected parts belonging to neighbouring pixels. Scattering predominantly effects neighboring pixels with very different phases . As the lens of the camera has to be focused to a fixed distance, e.g. 5m, the blurring effect favours scattering at distances which are shorter or longer than the one the camera is focused on. However, efforts and advances made by the manufacturers in order to minimize the influence of scattering raise hope that scattering will be overcome in the future. Mathematical models also help to minimize it (Mure-Dubois, 2007).

External superimposition seems to be the most dangerous error. It might very well ultimately prove to fundamentally limit the possibilities to use 3Dcameras for architectural recording. The reason for the occurrence of this effect lies in the fact that the array of diodes emitting the IR-Signal always illuminates the whole scene all at once. Small parts of diffusely reflected light from different parts of the object may superimpose the directly reflected signals on their way back to the camera. The possible influence of diffuse reflections on a wall, causing a phase shift of a directly reflected signal, can be depicted only schematically as in following figures:

21

Possible influence of diffuse reflection on direct measurement

External superimposition effecting a phase delay, respectively a distance shift

The impact of the effect is demonstrated in following figure. Appropriate measurements were made to a vertical edge, formed by two flexible planes. In the experiment the angle between the planes was modified while the position of the edge itself did not move. The smaller the angle between the planes, the more deformed the edge will appear. However, the distances measured more or less exactly into the centre of the corner, into the intersection of the planes, are those which are less deformed by the superimposition. The amount of error thus may reach considerable amounts.

22

Interaction between topography of the architecture and the measurement with a)d) increasing deformation with smaller angle; e),f) same angle but f) one plane less reflecting

The topography of an area can also mean the surface shape and features themselves or in a broader sense, topography is concerned with local detail in general. Here the feature and shapes of the reflecting surfaces are concerned, showing changes occurred in images as the position of the surfaces is changed. It can be easily observed that as the angle between the planes decreases the image gets more deformed keeping the position of edges constant.

23

6) Practical Use Of TOF Camera

The examples prove the wide range of possibilities inherent in the new methodology, which can be applied to a variety of tasks that currently require several different measuring devices, like the electronic-distance-meter,

tacheometer, laserscanner or camera. The possibilities of the new system are demonstrated in regards to distance-measurement, angle-measurement, taking images, measuring point clouds or using it like a video-camera at a high frame rate. The 2D- / 3D-system has all the fundamental capabilities of common contemporary measuring devices. Will it be possible to discard these instruments in favour of the 3D-camera in the future? Will it be possible to use it efficiently despite the handicap provided by external superimposition? The following experiments try to find some answers.

User interface of the 2D- / 3D-system

24

The 2D- / 3D-system in exchange to a tacheometer

With the tacheometer polar coordinates are measured in reference to the vertical: the zenith distance, the horizontal angle and the distance. In architectural surveying it is often used to establish a network or to capture single points. Whether the 3D-camera can replace a tacheometer is largely dependent on the orientation of the camera and on its accuracy. To prove this, the 2D- / 3D-camerasystem was mounted on a tacheometer tripod and leveled with a bubble-tube. The 3D camera was used for distance measurement, the 2D-camera for angle measurement because of its higher angle resolution. Different types of special targets were developed. E.g. in various sectors rods were marked with reflecting film or reflecting spheres.

Targets for automatic identification of single points

Although the cameras angle resolution is rather poor, distances could still be detected sufficiently well, especially as far as measurements to sphericalreflectors are concerned. In order to carry out angle-measurement well-defined points in the same line e.g. the reflecting spheres could be chosen automatically. Thus automated extraction and identification of reflecting points of high intensity was possible. Distances of more than 20m were reached, however with rapidly decreasing accuracy down to + 5cm. Faults caused by excessive intensity of the signal amplitude at the reflecting points had to be considered. If necessary, these strongly reflecting points can be used for identification purposes only, while less
25

reflecting adjacent (red) spheres may be utilized for angle measurement respectively for distance measurement. The latest generation cameras, however, should yield better results. It will also be easier to measure distances longer than half a wavelength, because increased switching times of the measuring frequency make it possible to determine n in the equation d = (n x + .) using different frequencies. The present means of interaction using a measuring rod is a typical modus operandi of the 3D-camera. It cannot be realized with other instruments this way. Only here different targets can be reached all at once. The eventual replacement of the tacheometer by the 3D-camera seems a likely future course for certain areas, e.g. for documenting the advancing process of measurements at an archaeological excavation site, or for topographical purposes at short distances. In these applications of measurements to reflecting targets faults caused by external superimposition may be sufficiently suppressed. It has also been proved that a network can easily be referenced by moving a GPS-antenna across the field of sight with the camera positioned on a tripod. Single-frame exposure time and GPS-coordinate measurement were synced.

Scanning with the 3D-camera

With the 3D-system being used as a scanner special attention should be paid to two different working modes: In the first one, the camera is mounted on a tripod to be turned around a fixed e.g. vertical - axis, whereas in the second one the camera may be moved freely around an object or moved about in a room. In both cases every single frame delivers a cloud of thousands of points that are fixed to the

26

camera system. In either of the two modes good starting points for matching are obtained by the procedure thanks to the high frame rate and the frames overlapping heavily. However, external superimposition may cause considerable deformations and will deteriorate the quality of matched point-clouds in any case. To minimize the influence the point-clouds of structures scanned with the 3Dsystem should preferably only be neighboring small reflecting areas. Figure gives an example of a suitable object. The point cloud was constructed from 50 frames, taken from only three positions with the camera on a tripod. This unfavourable way of recording can easily be identified by the characteristic structure of points forming rows.

27

Visualization

Photorealistic texturing requires geometry data and pre-rectified images. The images used for photorealistic modeling were taken with the 2D-camera and matched with the coordinates taken by the 3D-camera. Mapping and rectifying of the images was done here using the technology and the algorithms of photo tacheometry (Scherer, 2006). Figure shows an image of a three-dimensional work of art consisting of coloured planes facing each other in different angles. In this example no effects by external superimposition on the coordinates were to be expected. Changes in colour result from to the lower quality of the 2D video camera.

In the visualization of built structures the level of detail (LOD) is characterized by four degrees. LOD3 describes models of buildings that have been texturized on the outside. In these cases the general geometry recording resolution goes down to a few decimeters. LOD4 is reserved to describe the most detailed models, the interior of buildings as well as detailed textures. The visualization of the niche in example figure 12 shows that - in spite of superimposition-effects planes can be extracted sufficiently well to build a LOD3-model.
28

7) Technologies Developed for Time Of Flight Camera

Several different technologies for time-of-flight cameras have been developed. Pulsed light source with digital time counters There are devices with a pulsed laser and a custom imaging integrated circuit with a fast counter behind every pixel. These devices produce depth values for each pixel on every frame. Typical image sizes are 128 x 128 pixels. Ranges up to 22,000 feet with an eye-safe narrow beam have been achieved. Detectors are typically InGaAs (indium-gallium-arsenide) devices. RF-modulated light sources with phase detectors Photonic Mixer Devices (PMD) and the Swiss Ranger works by modulating the outgoing beam with an RF carrier, then measuring the phase shift of that carrier on the receive side. This is a compact, short-range device. This approach has a modular error challenge; ranges are mod the maximum range, which is the RF carrier wavelength. With phase unwrapping algorithms, the maximum uniqueness range can be increased. The Swiss Ranger has ranges of 5 or 10 meters, with 176 x 144 pixels. The PMD can provide ranges up to 60m. Illumination is pulsed LEDs, rather than a laser. The demodulation is usually achieved by gating the sensor in synchrony with the light source modulation, so in essence they are range gated imagers.

29

Range gated imagers This is the most promising technology, invented by Antonio Medina. The phase detector is the gate or shutter in the camera. The gate allows collection of portions S2 and S1 of the received light pulse S. The portions are dependent of the time of arrival, and range is derived from them according to Medina's equation, z =R(S2-S1)/2S+R/2 for an ideal camera. R is the camera range, determined by the round trip of the light pulse. The Z-cam and Canesta 3D cameras are range-gated systems. Similar principle is used in the ToF camera line developed by Fraunhofer Institute of Microelectronic Circuits and Systems and TriDiCam. These cameras employ photo detectors with fast electronic shutter used as a gated integrator and pulsed lasers. There are also "range gated imagers", which are not 3D cameras. The gate is open to collect the totality of the reflected pulse and nothing else. Anything outside a specified distance range can be suppressed. Those are useful for seeing through fog. A pulsed laser provides illumination, and an optical gate allows light to reach the imager only during the desired time period.

30

8)

Advantages of Time Of Flight Camera

Simplicity In contrast to stereo vision or triangulation systems, the whole system is very compact: the illumination is placed just next to the lens, whereas the other systems need a certain minimum base line. In contrast to laser scanning systems, no mechanical moving parts are needed. Efficient distance algorithm It is very easy to extract the distance information out of the output signals of the TOF sensor, therefore this task uses only a small amount of processing power, again in contrast to stereo vision, where complex correlation algorithms have to be implemented. After the distance data has been extracted, object detection, for example, is also easy to carry out because the algorithms are not disturbed by patterns on the object. Speed Time-of-flight cameras are able to measure the distances within a complete scene with one shot. As the cameras reach up to 100 frames per second, they are ideally suited to be used in real-time applications.

31

9)

Shortcomings in Time Of Flight Camera

Background light Although most of the background light coming from artificial lighting or the sun is suppressed, the pixel still has to provide a high dynamic range. The background light also generates electrons, which have to be stored. For example, the illumination units in today's TOF cameras can provide an illumination level of about 1 watt. The Sun has an illumination power of about 50 watts per square meter after the optical bandpass filter. Therefore, if the illuminated scene has a size of 1 square meter, the light from the sun is 50 times stronger than the modulated signal. Interference If several time-of-flight cameras are running at the same time, the cameras may disturb each others' measurements. There exist several possibilities for dealing with this problem:

Time multiplexing: A control system starts the measurement of the individual cameras consecutively, so that only one illumination unit is active at a time.

Different modulation frequencies: If the cameras modulate their light with different modulation frequencies, their light is collected in the other systems only as background illumination but does not disturb the distance measurement.

32

Multiple reflections In contrast to laser scanning systems, where only a single point is illuminated at once, the time-of-flight cameras illuminate a whole scene. Due to multiple reflections, the light may reach the objects along several paths and therefore, the measured distance may be greater than the true distance.

33

10) Applications Of TOF Camera

Automotive applications Time-of-flight cameras are also used in assistance and safety functions for advanced automotive applications such as active pedestrian safety, precrash detection and indoor applications like out-of-position (OOP) detection.

Human-machine interfaces / gaming

As time-of-flight cameras provide distance images in real time, it is easy to track movements of humans. This allows new interactions with consumer devices such as televisions. Another topic is to use this type of cameras to interact with games on video game consoles.

34

Measurement / machine vision

Other applications are measurement tasks, e.g. for the fill height in silos. In industrial machine vision, the time-of-flight camera helps to classify objects and help robots find the items, for instance on a conveyor. Door controls can distinguish easily between animals and humans reaching the door. Robotics Another use of these cameras is the field of robotics: Mobile robots can build up a map of their surroundings very quickly, enabling them to avoid obstacles or follow a leading person. As the distance calculation is simple, only little computational power is used.

35

CONCLUSION
On the one side, fundamental difficulties using 3D-cameras for architectural surveying are to be outlined, and on the other side possibilities for the use of a 2D- / 3D-system replacing traditional measuring instruments are to be shown.

The most important results from these first experiences are as follows:

- When modelling structures the geometrical accuracy depends on the structures themselves and on the geometrical relation between the camera and the object. The strong interdependence caused by external superimposition has to be taken into account.

- Geometrical correctness can often not be guaranteed.

- A sufficient resolution of point-clouds can be reached only with a large number of frames; the most efficient way to arrange them will have to be examined.

- The aforementioned method of recording with the 3D-camera is suited for recording and visualization with LOD3-quality.

- The combination of a 3D-camera with a 2D-camera of higher resolution yields improved coordinate quality.

36

- The large variety of possibilities to combine the different functionalities may allow to replace traditional measuring instruments and measuring methods under certain conditions.

Although many instrumental parameters will see additional improvements in the future, the external superimposition error is inherent in the technology itself. It may cause uncontrollable alterations in the distances measured. Determining accurate boundaries for widespread and common use of 3D-cameras in surveying might prove to be a long way.

37

BIBLIOGRAPHY

www.wikipedia.org/wiki www5.informatik.uni-erlangen.de/.../time-of-flight-tof-group www.mesa-imaging.ch/pdf/Application_SR3000_v1_1.pdf www.artts.eu/publications/oprisescu_et_al_ieee.pdf www.springerlink.com/index www.cs.unc.edu/~lguan/publications/guan_pollefeys_eccv08.pdf

38