3D Time-of-flight distance measurement with custom - Universität ...

More documents

Recommendations

Info

SUMMARY AND PERSPECTIVE 181 7. Summary and Perspective In this thesis we have presented the world’s first all-solid-state 3D time-of-flight range camera that is based on a customized photoASIC and does not require any mechanically moving parts. We achieve a distance resolution of a few centimeters over the range of several meters at a frame rate of 10 Hz for non-cooperative targets (reflectivity 0.7). The unambiguous distance range is 7.5 meters for 20 MHz modulation frequency and 20 meters for a modulation frequency of 7.5 MHz. We have given an overview of the existing range measurement techniques and compared them to the time-of-flight method. Optical interferometry mainly targets measurements in the range of several centimeters, and is therefore not a real competitor in the 1-100 m range. The other popular ranging method, triangulation, either active or passive, suffers from shadowing effects and requires a large triangulation base and hence a large overall system size for good accuracy and distance range. Furthermore, we have described and compared the classical CCD image sensors with CMOS APS imagers, which have gained more and more importance in recent years. One of the “secrets” of this CMOS APS boom is that they can be fabricated with standard technologies that are steadily being miniaturized and improved, driven by the computer industry. This makes it possible to implement more and more functionality within the sensor device and to realize more pixels on the same silicon area. From the existing range measurement techniques the time-of-flight method benefits most from this development in the semiconductor industry. While all range measurement systems will more or less decrease in size (note: triangulation will always require a large triangulation base for good accuracy), the TOF method will additionally gain in measurement resolution. This is mainly because the miniaturization in integrated circuits in general also leads to faster, lower noise circuits and therefore improves the time- and thus distance-resolution of TOF systems. For these reasons and in view of the additional advantages (no
182 CHAPTER 7 shadowing problems, compact setup) we are convinced that time-of-flight is the 3Dmeasurement technique of the future. To realize non-scanning 3D TOF cameras, an array of so-called demodulation pixels is required. Each of the pixels in this array must have the functionality to measure the arrival time of a modulated (optical) radiation field in addition to the intensity. The task of measuring the time of flight can be solved by synchronously sampling a periodically CW modulated optical signal. We have introduced such “sampling” demodulation pixels of different architectures. Our pixels have been implemented in a 2 µm CMOS/CCD technology. They make use of both the advantages of a CMOS-APS, such as random pixel access, and also the advantages of the CCD technique, which allows the addition of optically generated charge carriers, essentially free of noise, and their movement and distribution into defined storage areas in the semiconductor. We were able to show that, for a good homogenous performance of the pixels, the light sensitive area (photogate) should be accessed from only one (or two) site(s) since process inhomogeneities make it impossible to achieve homogenous sampling point sensitivity for a four side accessed photogate. Especially since we use active illumination with the TOF method, the optical fill factor of the pixels is of essential importance. The 1-tap pixel that we implemented has an optical fill factor of 100% for the line sensor realization and 20% for the realization as an image sensor. That is already competitive with CMOS APS sensors. Additionally, the simple geometry of the light sensitive areas (lines of light sensitive areas, i.e. 100% fill factor in one dimension and 20% in the other dimension) makes it easy to realize and assemble highly efficient (cylindrical) microlenses. For these simple structures microlenses can drastically improve the optical fill factor (up to 100% for collimated light), so that, compared to our current setup, either the optical power or the integration time can be reduced by a factor of five. (For other photodetector arrays with asymmetrical photosites an efficient fill factor improvement by microlenses is often a problem). We have extensively characterized and described the demodulation performance of the 1-tap pixel. For 20 MHz modulation frequency we measured a demodulation contrast of 40 % for a 630 nm modulated light source. With a simple model we could show that, even under DC conditions, the shutter efficiency of the pixel can never be better than 60 % to 70 %, since a fraction of light always travels directly
Page 1 and 2:
3D Time-of-flight distance measurem
Page 3 and 4:
To my parents
Page 5 and 6:
Meinen Eltern
Page 7 and 8:
II 4. Power budget and resolution l
Page 10 and 11:
Abstract Since we are living in a t
Page 12:
centimeter accuracy. With the singl
Page 15 and 16:
X Eine elegantere Methode zur Entfe
Page 18 and 19:
INTRODUCTION 1 1. Introduction One
Page 20 and 21:
INTRODUCTION 3 light source observe
Page 22 and 23:
INTRODUCTION 5 Propagation time, ho
Page 24 and 25:
INTRODUCTION 7 Chapter 3 gives a sh
Page 26 and 27:
OPTICAL TOF RANGE MEASUREMENT 9 2.
Page 28 and 29:
Page 30 and 31:
OPTICAL TOF RANGE MEASUREMENT 13 pr
Page 32 and 33:
OPTICAL TOF RANGE MEASUREMENT 15 hi
Page 34 and 35:
OPTICAL TOF RANGE MEASUREMENT 17 Pu
Page 36 and 37:
OPTICAL TOF RANGE MEASUREMENT 19 Ad
Page 38 and 39:
OPTICAL TOF RANGE MEASUREMENT 21 pr
Page 40 and 41:
OPTICAL TOF RANGE MEASUREMENT 23 wh
Page 42 and 43:
OPTICAL TOF RANGE MEASUREMENT 25 δ
Page 44 and 45:
OPTICAL TOF RANGE MEASUREMENT 27 c(
Page 46 and 47:
OPTICAL TOF RANGE MEASUREMENT 29 A
Page 48 and 49:
OPTICAL TOF RANGE MEASUREMENT 31 ph
Page 50 and 51:
OPTICAL TOF RANGE MEASUREMENT 33 Ts
Page 52 and 53:
OPTICAL TOF RANGE MEASUREMENT 35 f1
Page 54 and 55:
OPTICAL TOF RANGE MEASUREMENT 37 an
Page 56 and 57:
OPTICAL TOF RANGE MEASUREMENT 39 in
Page 58 and 59:
OPTICAL TOF RANGE MEASUREMENT 41 Ss
Page 60 and 61:
OPTICAL TOF RANGE MEASUREMENT 43 Fr
Page 62 and 63:
OPTICAL TOF RANGE MEASUREMENT 45 1
Page 64:
Page 67 and 68:
50 CHAPTER 3 the smearing effect, r
Page 69 and 70:
52 CHAPTER 3 3.1 Silicon properties
Page 71 and 72:
54 CHAPTER 3 (a) Figure 3.3 Optical
Page 73 and 74:
56 CHAPTER 3 Quantum efficiency in
Page 75 and 76:
58 CHAPTER 3 The different operatio
Page 77 and 78:
60 CHAPTER 3 (a) (b) Depletion widt
Page 79 and 80:
62 CHAPTER 3 Very special processes
Page 81 and 82:
64 CHAPTER 3 1 kT Dn = ⋅ vth ⋅
Page 83 and 84:
66 CHAPTER 3 The proportionality fa
Page 85 and 86:
68 CHAPTER 3 difference of only 1 V
Page 87 and 88:
70 CHAPTER 3 Flicker noise is cause
Page 89 and 90:
72 CHAPTER 3 A s = sf ⋅ q C ⎡ V
Page 91 and 92:
74 CHAPTER 3 amount of the charge p
Page 93 and 94:
76 CHAPTER 3 substrate (Equation 3.
Page 95 and 96:
78 CHAPTER 3 charge carriers the CT
Page 97 and 98:
80 CHAPTER 3 without causing short
Page 99 and 100:
82 CHAPTER 3 Over an address decode
Page 101 and 102:
84 CHAPTER 3 Orbit’s 2.0µm CMOS/
Page 103 and 104:
86 CHAPTER 4 Figure 4.1 P lens Lamb
Page 105 and 106:
88 CHAPTER 4 Required optical power
Page 107 and 108:
90 CHAPTER 4 4.2 Noise limitation o
Page 109 and 110:
92 CHAPTER 4 number of pseudo-backg
Page 111 and 112:
94 CHAPTER 4 Cmod = 1 QE(λ) = 0.65
Page 113 and 114:
96 CHAPTER 4 Range accuracy (std) i
Page 116 and 117:
DEMODULATION PIXELS IN CMOS/CCD 99
Page 118 and 119:
Page 120 and 121:
Page 122 and 123:
Page 124 and 125:
Page 126 and 127:
Page 128 and 129:
Page 130 and 131:
Page 132 and 133:
Page 134 and 135:
Page 136 and 137:
Page 138 and 139:
Page 140 and 141:
Page 142 and 143:
Page 144 and 145:
Page 146 and 147:
Page 148 and 149: DEMODULATION PIXELS IN CMOS/CCD 131
Page 168 and 169: IMAGING TOF RANGE CAMERAS 151 6. Im
Page 170 and 171: IMAGING TOF RANGE CAMERAS 153 contr
Page 172 and 173: IMAGING TOF RANGE CAMERAS 155 6.1.2
Page 174 and 175: IMAGING TOF RANGE CAMERAS 157 gate
Page 176 and 177: IMAGING TOF RANGE CAMERAS 159 6.2 2
Page 178 and 179: IMAGING TOF RANGE CAMERAS 161 n+ di
Page 180 and 181: IMAGING TOF RANGE CAMERAS 163 pixel
Page 182 and 183: IMAGING TOF RANGE CAMERAS 165 Power
Page 184 and 185: IMAGING TOF RANGE CAMERAS 167 Delay
Page 186 and 187: IMAGING TOF RANGE CAMERAS 169 6.3 3
Page 188 and 189: IMAGING TOF RANGE CAMERAS 171 Gate
Page 190 and 191: IMAGING TOF RANGE CAMERAS 173 Bound
Page 192 and 193: IMAGING TOF RANGE CAMERAS 175 Figur
Page 194 and 195: IMAGING TOF RANGE CAMERAS 177 Final
Page 196 and 197: IMAGING TOF RANGE CAMERAS 179 #5 #6
Page 200 and 201: SUMMARY AND PERSPECTIVE 183 into th
Page 202: SUMMARY AND PERSPECTIVE 185 or phas
Page 205 and 206: 188 CHAPTER 8 8.2 Typical parameter
Page 207 and 208: 190 CHAPTER 8 Spectral luminous int
Page 209 and 210: 192 CHAPTER 8 MCD03S: Conditions SC
Page 211 and 212: 194 CHAPTER 8 MCD06S: Measurement c
Page 213 and 214: 196 [BRK] Brockhaus, “Naturwissen
Page 215 and 216: 198 [KAI] I. Kaisto, et al., “Opt
Page 217 and 218: 200 [SP1] T. Spirig et al., “The
Page 220 and 221: ACKNOWLEDGMENTS 203 Acknowledgments
Page 222: 206 Publication list 1. R. Lange, P
show all

3D Time-of-flight distance measurement with custom - Universität ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?