Search

Panoramic annular lens (PAL) system consists of a panoramic head block unit and a relay lens system, adopting a plane cylindrical projection method to project a cylindrical field of view around the optical axis 360° into a two-dimensional planar annular region to realize super hemispherical imaging. PAL system requires large field of view imaging on a limited image plane, which inevitably results in lower image resolution and poor local detail imaging. However, the information processing speed gets faster with the development of 5G, which leads to a growing demand for better image quality and clarity. This paper presents a PAL system matching image sensors with 4K resolution, which can demonstrate the details about the object and get a smooth and ultra-high definition image. Unlike other sensors, it has large image faces and tiny pixels and the effective number of the pixel is about 24.3 million (4K × 6K). The proposed PAL system has 6 sets of 9 lenses with compact structure. The F-theta distortion of the full field of view is less than 1%, and the modulation transfer function value of each field of view is greater than 0.5 at the Nyquist spatial frequency of 130lp/mm which approaches diffraction limitation. The designed system has good imaging quality and meets the application requirements.

In recent years, with the vigorous development of artificial intelligence and autonomous driving technology, the importance of scene perception technology is increasing. Unsupervised deep learning based methods have demonstrated a certain level of robustness and accuracy in some challenging scenes. By inferring depth from a single input image without any ground truth label, a lot of time and resources can be saved. However, unsupervised depth estimation has defects in robustness and accuracy under complex environment which could be improved by modifying network structure and incorporating other modal information. In this paper, we propose an unsupervised, monocular depth estimation network achieving high speed and accuracy, and a learning framework with our depth estimation network to improve depth performance by incorporating transformed images across different modalities. The depth estimator is an encoder-decoder network to generate the multi-scale dense depth map. The sub-pixel convolutional layer is adopted to obtain depth super-resolution by replacing the up-sample branches. The cross-modal depth estimation using near-infrared image and RGB image enhances the performance of depth estimation than pure RGB image. The training mode is to transfer both images to the same modality and then carry out super-resolved depth estimation for each stereo camera pair. Compared with the initial results of depth estimation using only RGB images, the experiment verifies that our depth estimation network with the cross-modal fusion system designed in this paper achieves better performance on public datasets and a multi-modal dataset collected by our stereo vision sensor.

The Visually Impaired People (VIP) have the difficulty in perceiving the accurate localization in their daily life. Developing an efficient algorithm to address the localization issues of the VIP is crucial. Visual Place Recognition (VPR) refers to using the image retrieval algorithms to determine the location of a query image in the database, which is promising to help the VIP solve their localization problems. However, the accuracy of VPR is directly affected by the changes of scene appearances such as illumination, seasons and viewpoints. Therefore, finding a method to extract robust image descriptors under the changes of scene appearance is one of the most critical tasks in current VPR research. In this paper, we propose a VPR approach to assist the localization and navigation of visually impaired pedestrians. The core of our proposal is a combination of multi-level descriptors by using appropriate descriptors: the whole image, local regions and key-points, aimed to enhance the robustness of VPR. The matching procedure between query images and database images includes three steps. Firstly, we obtain the Convolutional Neural Networks (CNN) features of the whole images from a pre-trained GoogLeNet, and the Euclidean distances between the query images and the database images are computed to determine the top 10 good matches. Secondly, local salient regions are detected from the top-10 best-matched images with Non-Maximum Suppression (NMS) to control the number of bounding boxes. Thirdly, we detect the SIFT key-points and extract the geodesc descriptors of the key-points, from the local salient region, and determine the top 1 among the top 10 good matches. In order to verify our approach, a comprehensive set of experiments has been conducted on dataset with challenging environmental changes, such as the GardensPointWalking dataset.

For Visually Impaired People (VIP), it’s very difficult to perceive their surroundings. To address this problem, we propose a scene understanding system to aid VIP in indoor and outdoor environments. Semantic segmentation performance is generally sensitive to the environment and illumination changes, including the change between indoor and outdoor environments and the change across different weather conditions. Meanwhile, most existing methods have paid more attention on either the accuracy or the efficiency, instead of the balance between both of them. In the proposed system, the training dataset is preprocessed by using an illumination-invariant transformation to weaken the impact of illumination changes and improve the robustness of the semantic segmentation network. Regarding the structure of semantic segmentation network, the lightweight networks such as MobileNetV2 and ShuffleNet V2 are employed as the backbone of DeepLabv3+ to improve the accuracy with little increasing of computation, which is suitable for mobile assistance device. We evaluate the robustness of the segmentation model across different environments on the Gardens Point Walking dataset, and demonstrate the extremely positive effect of the illumination-invariant pre-transformation in challenging real-world domain. The network trained on computer achieves a relatively high accuracy on ADE20K relabeled into 20 classes. The frame rate of the proposed system is up to 83 FPS on a 1080Ti GPU.

Optics and light-based technology are becoming more and more important as a cross-field of new science and engineering, and the scope of optics education should be extended. Since 2018, in Zhejiang University a new wave of general education reform has started, and our course Light-based Science and Technologies and Human Civilization has become a common core course. In the last 3 years our students mainly come from science and engineering, only a few students come from social science. From now on we will offer this course to multidisciplinary students, especially the students of politics, economics, law and arts. In this paper we present our learning organization, especially the group discussion in multidisciplinary students. The students from different specialties are divided into one group. The discussion topics are about research methodology, famous arguments in optics history, optics in daily life and imagination of the future.

With the increasing demands of visually impaired people, developing assistive technology to help them travel effectively and safely has been a research hotspot. Red, Green, Blue and Depth (RGB-D) sensor has been widely used to help visually impaired people, but the detection and recognition of glass objects is still a challenge, considering the depth information of glass cannot be obtained correctly. In order to overcome the limitation, we put forward a method to detect glass objects in natural indoor scenes in this paper, which is based on the fusion of ultrasonic sensor and RGB-D sensor on a wearable prototype. Meanwhile, the erroneous depth map of glass object computed by the RGB-D sensor could also be densely recovered. In addition, under some special circumstances, such as facing a mirror or an obstacle within the minimum detectable range of the RGB-D sensor, we use a similar processing method to regain depth information in the invalid area of the original depth map. The experimental results show that the detection range and precision of the RGB-D sensor have been significantly improved with the aid of ultrasonic sensor. The proposed method is proved to be able to detect and recognize common glass obstacles for visually impaired people in real time, which is suitable for real-world indoor navigation assistance.

Visually Impaired (VI) people around the world have difficulties in socializing and traveling due to the limitation of traditional assistive tools. In recent years, practical assistance systems for scene text detection and recognition allow VI people to obtain text information from surrounding scenes. However, real-world scene text features complex background, low resolution, variable fonts as well as irregular arrangement which make it difficult to achieve robust scene text detection and recognition. In this paper, a scene text recognition system to help VI people is proposed. Firstly, we propose a high-performance neural network to detect and track objects, which is applied to specific scenes to obtain Regions of Interest (ROI). In order to achieve real-time detection, a light-weight deep neural network has been built using depth-wise separable convolutions that enables the system to be integrated into mobile devices with limited computational resources. Secondly, we train the neural network using the textural features to improve the precision of text detection. Our algorithm suppresses the effects of spatial transformation (including translation, scaling, rotation as well as other geometric transformations) based on the spatial transformer networks. Open-source optical character recognition (OCR) is used to train scene texts individually to improve the accuracy of text recognition. The interactive system eventually transfers the number and distance information of inbound buses to visually impaired people. Finally, a comprehensive set of experiments on several benchmark datasets demonstrates that our algorithm has achieved an extraordinary trade-off between precision and resource usage.

According to the data from the World Health Organization, 285 million people are estimated to be visually impaired worldwide, and 39 million are blind. It is very difficult for visually impaired people to perceive and avoid obstacles at a distance during their travelling. To address this problem, we propose a sensor fusion system, which combines the RGBDepth sensor and millimeter wave radar sensor, to detect the surrounding obstacles. The range and velocity of multiple obstacles are acquired by the millimeter wave radar based on the principle of frequency modulated continuous wave. The positions of the obstacles are verified by the RGB-Depth sensor based on the contour extraction and MeanShift algorithm. The data fusion algorithm based on particle filters obtains accurate state estimation by fusing RGB-Depth data with millimeter wave radar data. The experiment results show that multiple obstacles with different ranges and angles are successfully detected by the proposed system. The measurement uncertainties are reduced by the data fusion system, meanwhile the effective detectable range is expanded compared to the detection with only RGB-Depth sensor. Moreover, the measurement results are stable when the illumination varies. As a wearable prototype, the sensor fusion system has the characteristics of versatility, portability and cost-effectiveness, which is very suitable for blind navigation application.

Feature matching is at the base of many computer vision algorithms such as SLAM, which is a technology widely used in the area from intelligent vehicles (IV) to assistance for the visually impaired (VI). This article presents an improved detector and a novel semantic-visual descriptor, coined SORB (Semantic ORB), combining binary semantic labels and traditional ORB descriptor. Compared to the original ORB feature, the new SORB performs better in uniformity of distribution and accuracy of matching. We demonstrate it through experiments on some open source datasets and several real-world images obtained by RealSense.

In recent years, with development of computer vision and robotics, a wide variety of localization approaches have been proposed. However, it is still challenging to design a localization algorithm that performs well in both indoor and outdoor environment. In this paper, an algorithm that fuses camera, IMU, GPS, as well as digital compass is proposed to solve this problem. Our algorithm includes two phases: (1) the monocular RGB camera and IMU are fused together as a VIO that estimates the approximate orientation and position; (2) the absolute position and orientation measured by GPS and digital compass are merged with the position and orientation estimated in first phase to get a refined result in the world coordinate. A bag-of-word based algorithm is utilized to realize loop detection and relocalization. We also built a prototype and did two experiments to evaluate the effectiveness and robustness of the localization algorithm in both indoors and outdoors environment.

The goal of National University Students’ Optical-Science-Technology Competition (NUSOSTC) is to provide a nation-wide platform for students from the colleges and universities, which have majors in the field of optics and photonics, to communicate and learning each other. Meanwhile, it works on pushing forward the popularity of optoelectronic knowledge, cultivating the students' teamwork and innovation ability, promoting higher education personnel training mode and practice teaching reform, and then improving the quality of talent training. The founding, organizational structure development and overall organizational arrangements of NUSOSTC were introduced in this paper. Besides, the competition logo, theme, title, final date, numbers of participating universities, undertaking universities and cities of the five NUSOSTCs held during 2008 to 2016 and the progress had been made were given in detail.

Starting from 2015, a general education course named “Light-based science and technologies and human civilization” has been offered in Zhejiang University. We try to give a humanism view angle to observe optics and optical engineering, and combine them with the relationship of human and the nature, the development of human society and human health. In this course we introduce different historical periods of light-based science and technologies, the great optical researchers, the typical research methods, advantages, academic discussions and the relationship with human civilization. The relevant cross-fields of learning and Nobel Prize winners are also included. This course provides the students with the typical examples about how academic revolution influences the world development, and also with humanism sight which exceeds the range of science and technologies themselves.

The optimized curriculum of College of Optical Science and Engineering is accomplished at Zhejiang University, based on new trends from both research and industry. The curriculum includes general courses, foundation courses such as mathematics and physics, major core courses, laboratory courses and several module courses. Module courses include optical system designing, optical telecommunication, imaging and vision, electronics and computer science, optoelectronic sensing and metrology, optical mechanics and materials, basics and extension. These curricula reflect the direction of latest researches and relates closely with optoelectronics. Therefore, students may combine flexibly compulsory courses with elective courses, and establish the personalized curriculum of “optoelectronics + X”, according to their individual strengths and preferences.

In order to enhance the practical education and hands-on experience of optoelectronics and eliminate the overlapping contents that previously existed in the experiments section adhering to several different courses, a lab course of “Applied Optoelectronics Laboratory” has been established in the College of Optical Science and Engineering, Zhejiang University. The course consists of two sections, i.e., basic experiments and project design. In section 1, basic experiments provide hands-on experience with most of the fundamental concept taught in the corresponding courses. These basic experiments including the study of common light sources such as He-Ne laser, semiconductor laser and solid laser and LED; the testing and analysis of optical detectors based on effects of photovoltaic effect, photoconduction effect, photo emissive effect and array detectors. In section 2, the course encourages students to build a team and establish a stand-alone optical system to realize specific function by taking advantage of the basic knowledge learned from section 1. Through these measures, students acquired both basic knowledge and the practical application skills. Moreover, interest in science has been developed among students.

It is highly important for visually impaired people (VIP) to be aware of human beings around themselves, so correctly recognizing people in VIP assisting apparatus provide great convenience. However, in classical face recognition technology, faces used in training and prediction procedures are usually frontal, and the procedures of acquiring face images require subjects to get close to the camera so that frontal face and illumination guaranteed. Meanwhile, labels of faces are defined manually rather than automatically. Most of the time, labels belonging to different classes need to be input one by one. It prevents assisting application for VIP with these constraints in practice. In this article, a face recognition system under unconstrained environment is proposed. Specifically, it doesn’t require frontal pose or uniform illumination as required by previous algorithms. The attributes of this work lie in three aspects. First, a real time frontal-face synthesizing enhancement is implemented, and frontal faces help to increase recognition rate, which is proved with experiment results. Secondly, RGB-D camera plays a significant role in our system, from which both color and depth information are utilized to achieve real time face tracking which not only raises the detection rate but also gives an access to label faces automatically. Finally, we propose to use neural networks to train a face recognition system, and Principal Component Analysis (PCA) is applied to pre-refine the input data. This system is expected to provide convenient help for VIP to get familiar with others, and make an access for them to recognize people when the system is trained enough.

The atmospheric boundary layer can be directly influenced by the ground and it is closely related to human activities, so the detection and investigation of the atmospheric boundary layer is very important. Due to the abundant rainfall in the plum rain season in southern China, the atmospheric boundary layer height (ABLH) is very different from any other time of the year. Lidar is an active remote-sensing instrument, and the advantage of high spatial and temporal resolution makes it very suitable for the detection of the atmosphere. In this paper, a three-dimensional (3D) polarized lidar is introduced and the structure will be given in detail. Compared to traditional one-direction ground-based lidar, the pointing of the 3D scanning lidar is very flexible and can be adjusted to any direction within the up hemisphere (360 degrees by 90 degrees) in a very short time. The ABLH in the plum rain season (from June to July 2016) over Hangzhou area (30°16′ N, 120°07′ E) was observed and different derivation methods, such as the wavelet covariance method, the gradient method, and the profile fitting method were carried out and compared in detail. The results show that the wavelet covariance method exhibits better stability than the gradient method and better accuracy than the profile fitting method. This work brings a more flexible and accuracy way for the ABLH detection and will be of great importance to the atmospheric study during the plum rain season.

Cross-sensitivity is a crucial parameter since it detrimentally affect the performance of an accelerometer, especially for a high resolution accelerometer. In this paper, a suite of analytical and finite-elements-method (FEM) models for characterizing the mechanism and features of the cross-sensitivity of a single-axis MOEMS accelerometer composed of a diffraction grating and a micromachined mechanical sensing chip are presented, which have not been systematically investigated yet. The mechanism and phenomena of the cross-sensitivity of this type MOEMS accelerometer based on diffraction grating differ quite a lot from the traditional ones owing to the identical sensing principle. By analyzing the models, some ameliorations and the modified design are put forward to suppress the cross-sensitivity. The modified design, achieved by double sides etching on a specific double-substrate-layer silicon-on-insulator (SOI) wafer, is validated to have a far smaller cross-sensitivity compared with the design previously reported in the literature. Moreover, this design can suppress the cross-sensitivity dramatically without compromising the acceleration sensitivity and resolution.

The ultrahigh static displacement-acceleration sensitivity of a mechanical sensing chip is essential primarily for an ultrasensitive accelerometer. In this paper, an optimal design to implement to a single-axis MOEMS accelerometer consisting of a grating interferometry cavity and a micromachined sensing chip is presented. The micromachined sensing chip is composed of a proof mass along with its mechanical cantilever suspension and substrate. The dimensional parameters of the sensing chip, including the length, width, thickness and position of the cantilevers are evaluated and optimized both analytically and by finite-element-method (FEM) simulation to yield an unprecedented acceleration-displacement sensitivity. Compared with one of the most sensitive single-axis MOEMS accelerometers reported in the literature, the optimal mechanical design can yield a profound sensitivity improvement with an equal footprint area, specifically, 200% improvement in displacement-acceleration sensitivity with moderate resonant frequency and dynamic range. The modified design was microfabricated, packaged with the grating interferometry cavity and tested. The experimental results demonstrate that the MOEMS accelerometer with modified design can achieve the acceleration-displacement sensitivity of about 150μm/g and acceleration sensitivity of greater than 1500V/g, which validates the effectiveness of the optimal design.

The research and protection of the ocean ecosystem are key works to maintain the marine status and develop marine functions. However, human’s knowledge about the ocean is greatly limited. Now, in situ, acoustic and remote sensing methods have been applied in the research to understand and explore the ocean. Especially, the lidar is one outstanding remote sensing method for its high spatial and temporal resolution as well as the ability of the vertical detection. Highspectral- resolution lidar (HSRL) employs an ultra-narrow spectral filter to distinguish scattering signals between particles and water molecules without assuming a lidar ratio and obtains optical properties of the ocean with a high accuracy. Nevertheless, the complexity of the seawater causes variable marine optical properties, which gives huge potentiality to develop a HSRL working at different wavelengths in order to promote the inversion accuracy and increase the detection depth. The field-widened Michelson interferometer (FWMI), whose central transmittance can be tuned to any wavelength and field of view is large, can be employed as the HSRL spectral filter and solves problems that the operating wavelength of the iodine filter is fixed and the field of view of Fabry-Perot interferometer is small. The principle of the HSRL based on the FWMI designing for the ocean remote sensing will be presented in detail. In addition, the availability of the application of the FWMI influenced by the disturbance of the states of Brillouin scattering is analyzed and the preliminary theory shows that the HSRL instrument basing on FWMI could be employed in the marine remote sensing with a high accuracy.

Mie scattering theory was shown in this paper to be suitable for analyzing the forward scattered light intensity distribution of micro-sized air bubble defects in glass, shining by a monochrome laser with a wavelength of 532um. The scattered light was measured by a high definition CCD camera. The scattering process can be classified as uncorrelated single scattering according to the properties of optical media. After calculating and smoothing the gray value of split rings of picture, Chahine algorithm was applied to reverse the size of defects. This technique was accurate to within 5% for defects with radii of

This paper discusses the pulse signal of a novel opto-mechanical zeroth-order grating transducer based on an anomalous diffraction phenomenon, Wood’s type anomaly and its corresponding tolerance analysis. In this device, tiny changes in the displacement of the nanostructured grating elements lead to a dramatic increase or decrease of the optical reflection amplitude. With this special feature, this structure is ideal to measure very small displacement. Unexpectedly, the original sinusoidal signal of the device develops into a new signal form, i.e. pulse signal with the decrease of the air gap between two layers of gratings. Thus the sensitivity of the structure is improved 8 times higher, as the slope of the pulse signal, namely 2.5%/nm, i.e. 0.65dB/nm, is 8 times higher than that of the original signal form, namely 0.3%/nm, i.e. 0.03dB/nm. However, this device is very sensitive to parameters including wavelength, period, duty ratio, air gap as well as thickness of the gratings. Thus, in this paper the performance of the structures with different parameter settings is analyzed and optimized through rigorous coupled wavelength analysis (RCWA) and 3-D finite difference time domain (FDTD) method. All the calculated data enables us to apply the structure into fields required for different sensitivities with different values of grating parameters and thus broadens the further usage of such novel structure. In addition, a synthetic tolerance analysis of the pulse signal is conducted and indicates the possibility of achieving an actual device with the highest slope superior to 0.5%/nm is close to 85% and the possibility that the highest slope of an actual device falls in the interval ranging from 1.0%/nm to 2.0%/nm is 64%. All the simulated data enables us to get a better understanding of the tolerance of the pulse signal and a guidance of successful realization of an actual device.

In this paper, a method to detect internal pocks and bubbles of optical elements based on laser line source scanning is proposed. In dark field environment, a laser line source is used to illuminate from one side of the glass under test, a high-resolution CCD camera is used to take pictures in front of the glass sample. Images which contain information of defects are acquired through rough scanning and accurate scanning. Accurate three-dimensional coordinates of the internal defects are acquired after image processing, which descript the characteristic information of internal defects quantificationally. Compared with the microscope imaging measurement, this proposed detection of defects in optics based on laser line source scanning has a relative aberration smaller than 2%. In addition, the detection time is approximately reduced to 20 minutes from 1 hour dramatically. The analysis indicates that the error of the position of defects is much smaller than the size of them, which means the position of the defects can be acquired accurately by this approach.

In this paper, a high-accuracy calibration method for angular measurement of deformed and curved Moire patterns, based on template matching algorithm, is presented. We report a feasible and accurate method, based on Talbot interferometry and Moire deflectometry, to measure long focal-length lenses. Theoretical analysis indicates that the precision of this method is mainly influenced by the angle of Moire patterns. However, it’s difficult to obtain high-accuracy angle of Moire patterns, since the Moire patterns derived from experiment are constantly deformed or curved. We demonstrate a method, based on template matching algorithm, to calibrate deformed and curved Moire patterns, thus their angle can be calculated fast and accurately in sub-pixel domain. Numerical analysis and simulation prove that the method mentioned above demonstrates high precision and stability, and experiment results show that the accuracy of the long focal lengths measurement is improved obviously.

In this paper, a high-accuracy calibration method for errors resulted from aberration in long focal length measurement, is presented. Generally, Gaussian Equation is used for calculation without consideration of the errors caused by aberration. However, the errors are the key factor affecting the accuracy in the measurement system of a large aperture and long focal length lens. We creatively introduce an effective way to calibrate the errors, with detailed analysis of the long focal length measurement based on divergent light and Talbot interferometry. Aberration errors are simulated by Zemax. Then, we achieve auto-correction with the help of Visual C++ software and the experimental results reveal that the relative accuracy is better than 0.01%.By comparing modified values with experimental results obtained in knife-edge testing measurement, the proposed method is proved to be highly effective and reliable.

Though the technology of grating side-coupling is often applied in fields, such as coupling of light of single wavelength or narrow waveband, pump of fiber laser, integration of optical waveguide, its application for broadband coupling of visible spectrum is rarely studied. Sunlight can concentrate and output at the edge of waveguides by integrating sub-wavelength gratings with waveguides, making it a novel solar concentrator. In this paper, we simulated different grating structures with the finite-difference time-domain solution software (FDTD) to obtain the optimal structure design, since different grating structures feature different diffractive efficiencies. The result demonstrates that the structures mentioned above all feature good diffractive efficiencies in broadband wavelength, among which the blazing grating reaches the largest efficiency, namely 48.8%.This kind of sub-wavelength gratings feature integration of small size, which makes it promising in absorption of solar energy, such as lumination, photovoltaic cell, space melting, etc.

In the flat sub-aperture stitching test, the dominant e error comes from two dimensional translation stage that carries the flat under test, which contains tilt and position error i.e., the alignment error. In order to analyze the influence of alignment error on stitching precision, we use Zemax optical software to simulate a system to detect the phase of each sub-aperture of measured flat and add tilt and position errors and random noise to sub-apertures. The simulation model was utilized in this paper to evaluate the mechanical precision of the translation stage in order to meet a required stitching precision of 1/1000λ.

By analyzing the spectral domain’s phase information, the spectrally resolved white-light interferometry (SRWLI) is capable of obtaining the profile with a single frame of interferogram. However, only one-dimensional (1-D) surfaces can be tested with this technique, otherwise the interference patterns under adjacent wavelengths would overlap each other, which would make the whole interferogram hard to identify. We present here a 2-D SRWLI method which can be applied to measure narrow rectangle areas. Frequency comb is produced by means of using a F-P etalon to filter the broadband source. With the filtered frequency comb illumination, the interference patterns under adjacent wavelengths would be separated by a little distance, which enables us to obtain a 2-D profile with a small width. The experimental details of measurement of a step sample are discussed in this paper.

Traditional interferometric profilemeter suffers from phase ambiguity problem and sensitivity to environmental disturbance, thus preventing their applications for on-line surface inspections. We propose a new method to obtain the object surface's two-dimensional profile in a single shot using a dispersive interferometer. An air-spaced Fabry-Perot etalon was applied in order to decompose the light from the broadband source into discrete monochromatic constituents with equal wavelength interval. A blazed grating is implemented to effectively separate the interferograms of difference wavelength. As a result, the interference patterns of different wavelengths distributed separately on the CCD camera. By analyzing these patterns one can get the configuration of the original surface. A one-dimensional profiling experiment is carried out to test a surface with steps and the mean error of the result is below 0.2μm.

White-light interferometry is always an extremely useful and powerful tool for optical surface measurement. In this paper, a novelty method based on rapid positioning and fast surface measurement is proposed. The adjustment of optical path difference (OPD) is done automatically by a wavelength scanning method, and the processing of white-light interference patterns is accelerated by GPGPU which stands for General-Purpose computation on Graphics Processing Units, also known as GPU Computing. Graphics Processing Units are high-performance many-core processors in which the data are processed in parallel. As a result, the time we need to obtain a surface shape is reduced to ten percents compared with the conventional white-light interferometry. With these advantages, it is possible to measure a rough and areal surface in a short time.

Absolute interferometric testing of spherical surfaces is one of effective methods for ultra-precision testing, which can eliminate the systematic errors. In this paper, we discuss the absolute measurement of spherical surfaces and analyze the influence of errors existed in the organism comprehensively. To improve the testing precision, it is necessary to remove the alignment errors such as tilt and defocus. The traditional methods remove alignment errors by low-level approximation, so they may be not effective for the small F-number surface. To solve this problem, we propose a novel measurement with high-order approximation model. The simulation and experiment results show that: while the error between high-order approximation model and low-level approximate model is small on tilt, it is large on defocus, and the error will become larger when the F-number of testing surface becomes smaller; alignment errors can be well adjusted by our high-order approximation model.