Search

LiDAR-based 3D object detection and panoptic segmentation are two crucial tasks in the perception systems of autonomous vehicles and robots. In this paper, we propose All-in-One Perception Network (AOP-Net), a LiDAR-based multi-task framework that combines 3D object detection and panoptic segmentation. In this method, a dual-task 3D backbone is developed to extract both panoptic- and detection-level features from the input LiDAR point cloud. Also, a new 2D backbone that intertwines Multi-Layer Perceptron (MLP) and convolution layers is designed to further improve the detection task performance. Finally, a novel module is proposed to guide the detection head by recovering useful features discarded during down-sampling operations in the 3D backbone. This module leverages estimated instance segmentation masks to recover detailed information from each candidate object. The AOP-Net achieves state-of-the-art performance for published works on the nuScenes benchmark for both 3D object detection and panoptic segmentation tasks. Also, experiments show that our method easily adapts to and significantly improves the performance of any BEV-based 3D object detection method., Under review

Introducing short-range attractions in Brownian systems of monodisperse colloidal spheres can substantially impact their structures and consequently their optical transport and rheological properties. Here, for size-fractionated colloidal emulsions, we show that imposing an intermediate strength of attraction, well above but not much larger than thermal energy ($$\approx 5.6$$ ≈ 5.6 $$k_{\textrm{B}}T)$$ k B T ) , through micellar depletion leads to a striking notch in the measured inverse mean free path of optical transport, $$1/\ell ^*$$ 1 / ℓ ∗ , as a function of droplet volume fraction, $$\phi$$ ϕ . This notch, which appears between the hard-sphere glass transition, $$\phi _{\textrm{g}}$$ ϕ g , and maximal random jamming, $$\phi _{\textrm{MRJ}}$$ ϕ MRJ , implies the existence of a greater population of compact dense clusters of droplets, as compared to tenuous networks of droplets in strongly attractive emulsion gels. We extend a prior decorated core-shell network model for strongly attractive colloidal systems to include dense non-percolating clusters that do not contribute to shear rigidity. By constraining this extended model using the measured $$1/\ell ^*(\phi )$$ 1 / ℓ ∗ ( ϕ ) , we improve and expand the microrheological interpretation of diffusing wave spectroscopy (DWS) experiments made on attractive colloidal systems. Our measurements and modeling demonstrate richness and complexity in optical transport and shear rheological properties of dense, disordered colloidal systems having short-range intermediate attractions between moderately attractive glasses and strongly attractive gels.

Hybrid organic inorganic formate perovskites, AB(HCOO)$_3$, is a large family of compounds which exhibit variety of phase transitions and diverse properties. Some examples include (anti)ferroelectricity, ferroelasticity, (anti)ferromagnetism, and multiferroism. While many properties of these materials have already been characterized, we are not aware of any study that focuses on comprehensive property assessment of a large number of formate perovskites. Comparison of the materials property within the family is challenging due to systematic errors attributed to different techniques or the lack of data. For example, complete piezoelectric, dielectric and elastic tensors are not available. In this work, we utilize first-principles density functional theory based simulations to overcome these challenges and to report structural, mechanical, dielectric, piezoelectric, and ferroelectric properties for 29 formate perovskites. We find that these materials exhibit elastic stiffness in the range 0.5 to 127.0 GPa , highly anisotropic linear compressibility, including zero and even negative values; dielectric constants in the range 0.1 to 102.1; highly anisotropic piezoelectric response with the longitudinal values in the range 1.18 to 21.12 pC/N, and spontaneous polarizations in the range 0.2 to 7.8 $\mu$C/cm$^2$. Furthermore, we propose and computationally characterize a few formate perovskites, which have not been reported yet.

Wireless system designers use wireless coverage prediction tools to assess the coverage that can be achieved (and the potential interference that may result) when transmitters or base stations are placed at various sites throughout a region. Current methods for assessing wireless coverage prediction accuracy tend to be ad hoc, have limited statistical power, and offer only limited insight concerning the quality of the results. Comparing coverage predictions to field measurement data would at first seem to be the best option for evaluating prediction accuracy. However, measurement data tends to be both sparse and expensive, while delivering few insights concerning the causes of prediction errors. Further, commercial wireless system planning tools generally incorporate at least rudimentary techniques for comparing predictions to measured data, but do not generally offer options for comparing predictions made using different models to each other. This is surprising given that the majority of such tools give users the option of using alternative prediction models. Prediction accuracy depends in equal measure upon the sophistication of the technique used to predict coverage and the quality of the geospatial data that describes both the natural and man-made obstacles and structures in the region. Here, we propose that single-factor comparison of prediction maps obtained using the same prediction technique but different geospatial data, or the same geospatial data but different prediction techniques, can help designers determine whether the prediction technique or the geospatial data is the limiting factor, and take appropriate action. In order to demonstrate the feasibility of this approach, we predicted the wireless coverage provided by 28 GHz base stations mounted on traffic signal arms at two different intersections on the University of British Columbia campus using two different wireless coverage prediction models but the same geospatial data. We then investigated the relative power of spatial autocorrelation, spatial cross-correlation, difference maps, and scatter plots to provide insights into similarities and differences between the prediction maps. The results demonstrate the feasibility of this approach and prepare the way for further study and development of single-factor comparison of wireless coverage prediction maps.

We consider Bidder Selection Problem (BSP) in position auctions motivated by practical concerns of online advertising platforms. In this problem, the platform sells ad slots via an auction to a large pool of $n$ potential buyers with independent values drawn from known prior distributions. The seller can only invite a fraction of $k

A static full-Stokes Fourier transform imaging spectropolarimeter incorporating a liquid-crystal polarization modulator (LPM) and birefringent shearing interferometer (BSI) is reported. It can decode the polarization information at each wavelength along the spatial dimension of a two-dimensional data array. The LPM has a high-speed time-division architecture and employs two ferroelectric liquid crystals and two wave plates to produce four polarization states, providing full-Stokes polarimetric information with a high signal-to-noise ratio. The BSI comprises two birefringent crystal plates and generates an optical path difference with good linear distribution for broadband interference, allowing a fast and high-precision spectral recovery. The optimized design of LPM and BSI are introduced in detail. Subsequently, the signal reconstruction is verified through simulations and experiments. The proposed scheme is highly efficient, exhibits a higher spectral resolution, and constitutes a compact technical approach to realize high-dimensional optical measurement.

A fast and accurate panoptic segmentation system for LiDAR point clouds is crucial for autonomous driving vehicles to understand the surrounding objects and scenes. Existing approaches usually rely on proposals or clustering to segment foreground instances. As a result, they struggle to achieve real-time performance. In this paper, we propose a novel real-time end-to-end panoptic segmentation network for LiDAR point clouds, called CPSeg. In particular, CPSeg comprises a shared encoder, a dual-decoder, and a cluster-free instance segmentation head, which is able to dynamically pillarize foreground points according to the learned embedding. Then, it acquires instance labels by finding connected pillars with a pairwise embedding comparison. Thus, the conventional proposal-based or clustering-based instance segmentation is transformed into a binary segmentation problem on the pairwise embedding comparison matrix. To help the network regress instance embedding, a fast and deterministic depth completion algorithm is proposed to calculate the surface normal of each point cloud in real-time. The proposed method is benchmarked on two large-scale autonomous driving datasets: SemanticKITTI and nuScenes. Notably, extensive experimental results show that CPSeg achieves state-of-the-art results among real-time approaches on both datasets., Comment: Accepted at ICRA 2023

An iterative local Fourier transform (ILFT)-based high-accuracy wavelength calibration for Fourier transform imaging spectrometer (FTIS) is proposed. The wavelength calibration for FTIS is to determine the relation between the wavelength and the wavenumber position. However, the wavenumber position solved by conventional method is only accurate up to integers restricted by the picket-fence effect of discrete Fourier transform. While the proposed ILFT can increase the accuracy of calculating the wavenumber position by combining the local Fourier transform and a few iterations. In this paper, the method is investigated in theory and then by simulations and experiments. The simulations show that the accuracy of the wavenumber position calculated by the ILFT is increased by 100 times than conventional method with noise, phase error, and non-uniform sampling of optical path difference. And the experimental results indicate that the ILFT decreases the absolute error of wavelength calibration from about 2.03 nm to 0.16 nm. Therefore, the method provides theoretical and technical support for FTIS and promotes the development of superior resolutions therein.

Fourier Transform Imaging Spectrometer(FTIS) is an instrument cable of acquiring two-dimensional spatial information and one-dimensional spectral information. The FTIS has attracted much attention and is widely applied in the fields like military reconnaissance, remote sensing, biomedicine, environmental monitoring, etc. The FTIS acquires the spectral intensity in different wavelengths by performing Fourier transform on the white light interference signal of the target generated by the FTIS. The spectral curve obtained directly by Fourier transform reflects the relationship between the wavenumber order and the spectral intensity. So wavelength calibration is required to convert the above relationship into the relationship between the wavelength and the spectral intensity, which makes it more intuitive. Therefore, wavelength calibration is a necessary step for FTIS to recovery spectrum. The traditional wavelength calibration method can only get the wavenumber order in the range of integer because of the picket fence effect in Fourier transform. It will definitely ignore the fractional part which results in the inaccurate wavenumber order, which will directly affect the precision of the wavelength calibration result. In order to solve this problem, a high-precision wavelength calibration method based on Fourier transform imaging spectrometer is proposed according to the principle of FTIS and Fourier transform. This method can calculate the wavenumber order with the precision of percentile, which will reduce the error of wavelength calibration effectively. As a result, the precision of spectral calibration can be increased eventually. This method realizes high-precision wavelength calibration by the way of adding zero to the interference fringe in the spatial domain. The core of the method is getting a more precise wavenumber order. The brief process of obtaining wavenumber order is as follows: First, the FTIS acquires the interference fringe of a monochromatic laser. Second, the original interference fringe is extrapolated with zero. Third, the subtle spectrum can be obtained by performing Fourier transform on the extrapolated interference fringe. Finally, the precise wavenumber order is calculated by dividing the abscissa of the peak value by the extrapolation multiple. The principle of this method is investigated and related simulations are then carried out. The simulation results indicate that the wavenumber order calculated by the method have the same precision with the preset parameters, which illustrates that the method can calculate the wavenumber order more accurately. Therefore, the method can improve the precision of the spectral calibration. Besides, related experiments are also performed. The laser interference fringes of different wavelengths generated by the actual FTIS all apply the method to get the wavenumber orders in the frequency domain. Then a curve which is the wavelength calibration function is fitted using the discrete relation between the wavelengths and the wavenumber orders. A laser whose wavelength is known is measured by the FTIS with the wavelength calibration function got by the proposed method. The error of the wavelength measurement result is one-fifth of the traditional method. The simulations and the experiment results indicate that the proposed method can improve the precision of the wavelength calibration, which provides the theory and technology support for spectral measurement using FTIS. It also provides a possibility for the development of FTIS towards the super resolution direction.

This study proposes a polychromatic interferometric illumination and polarimetric sensor-based imaging method for spectrum, polarization, and 3D shape, which are significant physical parameters of feature analysis for target detection. 5D-fusion sensing refers to the joint detection and fusion of the above 5D information, which is currently a great challenge. The method generates a polychromatic interference pattern using a Sagnac lateral shearing interferometer and projects it to the target. Then, interferograms modulated by the target are acquired during scanning. Fast Fourier transform (FFT) is performed on the interferograms to obtain their frequency spectra. The spectral and polarization information is extracted from the moduli of the frequency spectra. The 3D shape is recovered from the phase of the frequency spectra using the calibration data. The theory of 5D-fusion sensing is investigated, and verification experiments are then performed. The experiments indicate that the proposed method can fulfill 5D-fusion sensing in one scanning and with FFT using only one device compared with other separate methods. Consequently, the proposed method can improve the sensing and recognition ability of optical imaging technology, which provides great application potential in biomedicine, food safety, material analysis, criminal investigation, archeology, and other fields.

A dual-channel interferometer (DCI) is proposed for vibration-resistant optical measurement. The DCI employs a dual-channel interference optical path design, wherein an assistant channel is used to generate dense fringe patterns by introducing a spatial carrier frequency, and the coefficients of the vibration tilt phase plane are then extracted. The sparse fringe patterns of the main channel are combined with the coefficients of the tilt phase plane to calculate the measured phase distribution. This study reports the optical path and principles of the interference system. Simulations are performed using the proposed DCI phase extraction algorithm. To verify the DCI performance, phase measurements are carried out in a vibrating environment, and the results obtained using the proposed method are compared with those obtained using spatial phase-shifting interferometry (SPSI). The experimental results demonstrate that the DCI can effectively overcome the effect of vibration. Unlike spatial carrier frequency interferometry, although the spatial carrier frequency is also introduced in the assistant channel in the proposed DCI method, the dual-channel information extraction algorithm helps eliminate the calibration process of the retrace error before measurement, making this a calibration-free measurement. Moreover, the DCI utilizes environmental vibration to generate a ‘phase shift’, thus not only eliminating the need for a vibration isolation platform but also the need for a precise phase shifting device, with advantages such as structural simplicity, ease of operation, and low cost.

A new Fourier transform imaging spectrometer based on a focal plane birefringent interferometer (FPBI) is presented. The FPBI, located in front of the detector, is capable of performing spectral imaging measurements. It mainly consists of a birefringent plate and a birefringent wedge. The ordinary and extraordinary rays—with an optical path difference—are split by the FPBI and interfere on the focal plane. The spectral image of the target can be acquired via scene scanning and spectral recovery. The principle of interferometric imaging of the FPBI is investigated, and verification experiments are then performed. The experiments indicate that the FPBI not only provides effective spectral imaging measurements, but also presents the advantages of being ultra-compact and lightweight. As a result, it can be effectively applied in situations such as outdoor surveillance and airborne remote sensing.

A compact birefringent interferometer (CBI) for Fourier transform hyperspectral imaging is presented. The CBI employs only two birefringent crystal plates: a shearing plate (SP) and a compensation plate (CP). The SP generates the optical path difference (OPD) associated with the field of view for broadband interference. The CP compensates the constant term and square term OPDs of the SP to adjust the position of the zero-order fringe pattern and suppress inconsistent total OPDs and other nonlinear OPDs. This paper details the theoretically deduced OPDs and then presents simulation analyses and verification experiments conducted to investigate the OPD distribution characteristics. To verify the CBI performance, experimental spectral measurements and hyperspectral imaging were performed. The experimental results demonstrate that the CBI can suppress inconsistent total OPDs and other nonlinear OPDs with only two birefringent crystal plates, and therefore offers much promise for miniature and high-precision Fourier transform hyperspectral imaging.

China’s growth and development have opened the door to a new world. Shandong province’s 90 million inhabitants are entering into a consumption society and the waste stream from households, restaurants and commercials has become a challenge. So far, the waste has mostly been burned in backyards, thrown into rivers, put on open dumps or taken to landfills. The environmental consequence is strong. This study was carried out in Shandong province and presents the current situation of the municipal waste management. The result of the study is organised into social, economical, technical and environmental parameters. It mostly covers the years 2006 to 2010. In the discussion, the strength, weakness, opportunities, and threats to the management are analysed, which gives an overview of the complex situation. The final conclusion is that there are a lot of opportunities in developing municipal solid waste management in Shandong province since the work and planning is new and economy is good. Threats are for example a larger waste stream. The municipal waste management has some strengths, such as a lot of projects going on, but also a lot of weakness for instance implementation of the regulations and laws. Kinas starka tillväxt håller på att skapa ett konsumtionssamhälle som påminner om det västerländska. Med högre konsumtion och starkare ekonomi växer avfallsmängderna snabbt. I Shandong provinsen finns mer än 90 miljoner invånare som varje dag gör sig av med 0,5 – 1,5 kg sopor. Det är en utmaning för samhällsplanerana och ingenjörerna att ta hand om dessa volymer. Hittills har det mesta av avfallet bränts på bakgårdar, kastas i vattendrag, ut i naturen eller lagts på deponi. Miljöskadorna är stora. Denna studie beskriver avfallshanteringens situtation i Shandong provinsen. Faktan är uppdelad utefter sociala, ekonomiska, tekniska och ekologiska parametrar. Studien täcker åren 2006 till 2010. Olika avfallsanläggningar och konstruktioner besöktes, utländsk litterlatur användes och en handledare från ett kinesiskt provinskonstruktionsföretag bidrog med information och data. Studien är delvis dokumenterad med foton. I diskussionen analysers styrkor, svagheter, möjligheter och hot med avfallshanteringen för överblick av situationen. I slutsatsen framgår det att det finns många möjligheter i utvecklingen av ett avfallshanteringssystem eftersom det är nytt, flexibelt och ekonomin är god. Det finns även hårda miljö och säkerheteslagar som påminner om EUs och USAs. Till hoten räknas stora avfallsvolymer som kan vara svårkontrollerade, samt att alldeles för hårda lagar kan hindra nya innovationer och teknik. Styrkor som finns är att fler än 111 stycken projekt är på gång, och att kunskap finns på högre ledningsnivå. Svageheterna är dålig implementering av lagar samt dålig kunskap på lägre nivå. Eftersom avfallhanteringen är i ett initierande stadie så kan de val som görs nu påverka på väldigt långt sikt.