Fu-Pen Chiang

High-speed 3-D shape measurement based on digital fringe projection

A high-speed 3-D shape measurement technique based on digital fringe projection has been developed and experimented. This technique uses a computer-generated color fringe pattern whose red, green, and blue channels are sinusoidal fringe patterns with a 120-deg phase shift between neighboring channels. When this color fringe pat- tern is sent to a digital-micromirror-device (DMD) based video projector with no color filter, three gray-scale fringe patterns are repeatedly pro- jected to an object surface in sequence with a cycle time of approxi- mately 10 ms. The three phase-shifted fringe patterns deformed by the object surface are captured by a CCD camera with proper synchroniza- tion between the camera and the projector. The 3-D shape of the object surface is reconstructed by using a phase wrapping and unwrapping algorithm and a phase-coordinate conversion algorithm. Experimental results demonstrated the feasibility of this technique for high-speed 3-D shape measurement with a potential measurement speed up to 100 Hz.

Color-encoded digital fringe projection technique for high-speed three-dimensional surface contouring

A color-encoded digital fringe projection technique is proposed for high-speed 3-D surface contouring applications. in this technique, a color fringe pattern whose RGB components comprise three phase-shifted fringe patterns is created by software on a computer screen and then projected to an object by a novel computer-controlled digital projection system. The image of the object is captured by a digital camera positioned at an angle different from that of the projection system. The image is then separated into its RGB components, creating three phase-shifted images of the object. These three images are used to retrieve the 3-D surface contour of the object through the use of a phase wrapping and unwrapping algorithm. Only one image of the object is required to obtain the 3-D surface contour of the object. Thus contouring speed, limited only by the frame rate of the camera, can be dramatically increased as compared to that of the traditional phase-shifting techniques. The technique is especially useful in applications where the object being contoured is going through quasi-static or dynamic changes. The principle of the technique is described and some preliminary experimental results are presented.

Wavelet-based pavement distress detection and evaluation

An automated pavement inspection system consists of image acquisition and distress image processing. The former is accomplished with imaging sensors, such as video cameras and photomultiplier tubes. The latter includes distress detection, isolation, classification, evaluation, segmentation, and compression. We focus on wavelet-based distress detection, isolation, and evaluation. After a pavement image is decomposed into different-frequency subbands by the wavelet transform, distresses are transformed into high-amplitude wavelet coefficients and noise is transformed into low-amplitude wavelet coefficients, both in the high-frequency subbands, referred to as details. Background is transformed into wavelet coefficients in a low-frequency subband, referred to as approximation. First, several statistical criteria are developed for distress detection and isolation, which include the high-amplitude wavelet coefficient percentage (HAWCP), the high-frequency energy percentage (HFEP), and the standard deviation (STD). These criteria are tested on hundreds of pavement images differing by type, severity, and extent of distress. Experimental results demonstrate that the proposed criteria are reliable for distress detection and isolation and that real-time distress detection and screening is currently feasible. A norm for pavement distress quantification, which is defined as the product of HAWCP and HFEP, is also proposed. Experimental results show that the norm is a useful index for pavement distress evaluation.

Calibration of a three-dimensional shape measurement system

A 3-D surface shape measurement system based on a digital fringe projection and phase shifting technique is described. In this system, three phase-shifted fringe patterns and a centerline pattern are used to determine the absolute phase map of the object. This phase map is then converted to the x , y , and z coordinates of the object surface by a conversion algorithm. To determine the accurate values of the system parameters as required by the conversion algorithm, a two-step calibration procedure was developed. The parameters were first measured to determine their approximate values, then a calibration plate was measured by the system at various positions, and an iteration algorithm used to estimate the system parameters. Measurement results of several objects are presented. The standard deviation of the measurement error was found to be 0.23 mm.

Color phase-shifting technique for three-dimensional shape measurement

A color phase-shifting technique has been recently developed for high-speed 3-D shape measurement. In this technique, three sinusoidal phase-shifted images used for a measurement cycle in a traditional grayscale phase-shifting technique are encoded into one color image. Therefore, only a single color image is needed for reconstructing the 3-D surface shape of an object. The measurement speed can then be increased up to the frame rate of the camera. However, previous experimental results showed that the measurement accuracy of this technique was initially low, due largely to the coupling and imbalance of color channels. In this paper, two solutions, one software-based and one hardware-based, are proposed to compensate for these errors. Experimental results show that the second solution—modification of the camera together with an imbalance compensation algorithm—would effectively reduce the errors and produce better measurement results than the software-based compensation method. This technique has many potential applications in high-speed measurement, such as highway inspection and dynamic measurement of human body.

Color-coded binary fringe projection technique for 3-D shape measurement

Color coding has been used for 3-D shape measurement in many recently developed fringe projection techniques. Use of color allows for more information to be coded in the same number of patterns as compared to the black-and-white techniques. However, one major problem of using color is that the appearance of the color fringe patterns projected onto the object can be affected by the color of the object surface itself. Thus, correctly decoding the fringe patterns can be difficult and sometimes even impossible. We describe a color-coded binary fringe projection technique that solves this problem. The use of an adaptive threshold scheme enables the extraction of the 3-D information and texture of an object without being affected by the color of the object surface. The development of a color gray-code concept, which is an extension of the gray-code technique, further reduces decoding errors. In addition, this technique can be used to measure objects with discontinuous features. The system has small digitizing errors and its measurement accuracy is hardly affected by system noise and nonlinearity errors. The system setup, color pattern design, shape reconstruction, and experimental results are presented

Double three-step phase-shifting algorithm

We describe what we believe is a new phase-shifting algorithm called a double three-step algorithm developed to reduce the measurement error of a three-dimensional shape-measurement system, which is based on digital fringe-projection and phase-shifting techniques. After comparing the performance of different existing phase-shifting algorithms, we present the new double three-step algorithm based on the error analysis of the standard three-step algorithm. In this algorithm, three-step phase shifting is done twice with an initial phase offset of 60 degrees between them, and the two obtained phase maps are averaged to generate the final phase map. Both theoretical and experimental results showed that this new algorithm worked well in significantly reducing the measurement error.

Microscopic phase-shifting profilometry based on digital micromirror device technology.

A microscopic three-dimensional (3-D) shape measurement system based on digital fringe projection has been developed and experimentally investigated. A Digital Micromirror Device along with its illumination optics is integrated into a stereomicroscope, which projects computer-generated fringe patterns with a sinusoidal intensity profile through the microscope objective onto the object surface being measured. The fringe patterns deformed by the object surface are recorded by a CCD camera. The microscopic 3-D shape of the object surface is measured and reconstructed by use of a phase-shifting technique. We discuss design considerations and error analysis of the system. Experimental results successfully demonstrate the capability of this technique for surface profile measurement of rough surfaces at the micrometer level.

Trapezoidal phase-shifting method for three-dimensional shape measurement

We propose a novel structured light method, namely a trap- ezoidal phase-shifting method, for 3-D shape measurement. This method uses three patterns coded with phase-shifted, trapezoidal- shaped gray levels. The 3-D information of the object is extracted by direct calculation of an intensity ratio. Compared to traditional intensity- ratio-based methods, the vertical or depth resolution is six times better. Also, this new method is significantly less sensitive to the defocusing effect of the captured images, which makes large-depth 3-D shape mea- surement possible. If compared to sinusoidal phase-shifting methods, the resolution is similar, but the data processing speed is at least 4.5 times faster. The feasibility of this method is demonstrated in a previ- ously developed real-time 3-D shape measurement system. The recon- structed 3-D results show similar quality to those obtained by the sinu- soidal phase-shifting method. However, since the data processing speed is much faster 4.6 ms per frame, both image acquisition and 3-D re- construction can be done in real time at a frame rate of 40 fps and a resolution of 532500 points. This real-time capability allows us to mea- sure dynamically changing objects, such as human faces. The potential applications of this new method include industrial inspection, reverse engineering, robotic vision, computer graphics, medical diagnosis, etc.

Error compensation for a three-dimensional shape measurement system

We present an error compensation method for a full-field 3-D shape measurement system based on a digital fringe projection and phase shifting technique. The error map of the system is first established by comparing the measured coordinates with the coordinates defined by a coordinates measuring machine (CMM) at preselected sample points within the measurement volume. An eight-point interpolation algorithm based on the Shepards method is then used to compensate for the errors in the measured coordinates. Experimental results showed that the accuracy of the system was improved by more than 60% after error compensation.