Weiji He

Fast Depth Imaging Denoising With the Temporal Correlation of Photons

This paper proposes a novel method to filter out the false alarm of LiDAR system by using the temporal correlation of target reflected photons. Because of the inevitable noise, which is due to background light and dark counts of the detector, the depth imaging of LiDAR system exists a large estimation error. Our method combines the Poisson statistical model with the different distribution feature of signal and noise in the time axis. Due to selecting a proper threshold, our method can effectively filter out the false alarm of system and use the ToFs of detected signal photons to rebuild the depth image of the scene. The experimental results reveal that by our method it can fast distinguish the distance between two close objects, which is confused due to the high background noise, and acquire the accurate depth image of the scene. Our method need not increase the complexity of the system and is useful in power-limited depth imaging.

Adaptive Depth Imaging With Single-Photon Detectors

For active optical imaging, the use of single-photon detectors can greatly improve the detection sensitivity of the system. However, the traditional maximum-likelihood based imaging method needs a long acquisition time to capture clear 3-D image at low light levels. To tackle this problem, we present a novel imaging method for depth estimate, which can obtain the accurate 3-D image in a short acquisition time. Our method combines the photon-count statistics with the temporal correlations of the reflected signal. According to the characteristics of the target surface, including the surface reflectivity, our method is capable of adaptively changing the dwell time in each pixel. The experimental results demonstrate that the proposed method can quickly obtain the accurate depth image despite the existence of strong background noise.

Colored adaptive compressed imaging with a single photodiode.

Computational ghost imaging is commonly used to reconstruct grayscale images. Currently, however, there is little research aimed at reconstructing color images. In this paper, we theoretically and experimentally demonstrate a colored adaptive compressed imaging method. Benefiting from imaging in YUV color space, the proposed method adequately exploits the sparsity of the U, V components in the wavelet domain, the interdependence between luminance and chrominance, and human visual characteristics. The simulation and experimental results show that our method greatly reduces the measurements required and offers better image quality compared to recovering the red (R), green (G), and blue (B) components separately in RGB color space. As the application of a single photodiode increases, our method shows great potential in many fields.

Adaptive compressed photon counting 3D imaging based on wavelet trees and depth map sparse representation

We demonstrate a photon counting 3D imaging system with short-pulsed structured illumination and a single-pixel photon counting detector. The proposed multiresolution photon counting 3D imaging technique acquires a high-resolution 3D image from a coarse image and details at successfully finer resolution sampled along the wavelet trees and their depth map sparse representations. Both the required measurements and the reconstruction time can be significant reduced, which makes the proposed technique suitable for scenes with high spatial resolution. The experimental results indicate that both the reflectivity and depth map of a scene at resolutions up to 512×512 pixels can be acquired and retrieved with practical times as low as 17.5 seconds. In addition, we demonstrate that this technique has ability to image in presence of partially-transmissive occluders, and to directly acquire novelty images to find changes in a scene.

3D compressive spectral integral imaging.

A novel compressive 3D imaging spectrometer based on the coded aperture snapshot spectral imager (CASSI) is proposed. By inserting a microlens array (MLA) into the CASSI system, one can capture spectral data of 3D objects in a single snapshot without requiring 3D scanning. The 3D spatio-spectral sensing phenomena is modelled by computational integral imaging in tandem with compressive coded aperture spectral imaging. A set of focal stack images is reconstructed from a single compressive measurement, and presented as images focused on different depth planes where the objects are located. The proposed optical system is demonstrated with simulations and experimental results.

Speckle-Shifting Ghost Imaging

In this paper, we introduce speckle-shifting ghost imaging (SSGI) which uses several corresponding shifted groups of speckle patterns instead of random speckle patterns in “computational ghost imaging” (CGI) to improve the performance of edge detection. The shifting of speckle patterns makes SSGI directly achieve the edges of an unknown object without the clear “ghost” images. Numerical simulations and experiments are performed. It is seen that SSGI is applicable for both binary and gray-scale objects in noisy environments. This provides a great opportunity to pave the way for the real applications of CGI in remote sensing and biological imaging.

Three-dimensional Active Imaging using Compressed Gating

Due to the numerous applications employed 3D data such as target detection and recognition, three-dimensional (3D) active imaging draws great interest recently. Employing a pulsed laser as the illumination source and an intensified sensor as the image sensor, the 3D active imaging method emits and then records laser pulses to infer the distance between the target and the sensor. One of the limitations of the 3D active imaging is that acquiring depth map with high depth resolution requires a full range sweep, as well as a large number of detections, which limits the detection speed. In this work, a compressed gating method combining the 3D active imaging and compressive sensing (CS) is proposed on the basis of the random gating method to achieve the depth map reconstruction from a significantly reduced number of detections. Employing random sequences to control the sensor gate, this method estimates the distance and reconstructs the depth map in the framework of CS. A simulation was carried out to estimate the performance of the proposed method. A scene generated by the 3ds Max was employed as target and a reconstruction algorithm was used to recover the depth map in the simulation. The simulation results have shown that the proposed method can reconstruct the depth map with slight reconstruction error using as low as 7% detections that the conventional method requires and achieve perfect reconstruction from about 10% detections under the same depth resolution. It has also indicated that the number of detections required is affected by depth resolution, noise generated by a variety of reasons and complexity of the target scene. According to the simulation results, the compressed gating method is able to be used in the case of long range with high depth resolution and robust to various types of noise. In addition, the method is able to be used for multiple-return signals measurement without increase in the number of detections.

Range walk error correction using prior modeling in photon counting 3D imaging lidar

A real-time correction method for range walk error in photon counting 3D imaging Lidar is proposed in this paper. We establish the photon detection model and pulse output delay model for GmAPD, which indicates that range walk error in photon counting 3D imaging Lidar is mainly effected by the number of photons during laser echo pulse. A measurable variable – laser pulse response rate is defined as a substitute of the number of photons during laser echo pulse, and the expression of the range walk error with respect to the laser pulse response rate is obtained using priori calibration. By recording photon arrival time distribution, the measurement error of unknown targets is predicted using established range walk error function and the range walk error compensated image is got. Thus real-time correction of the measurement error in photon counting 3D imaging Lidar is implemented. The experimental results show that the range walks error caused by the difference in reflected energy of the target can be effectively avoided without increasing the complexity of photon counting 3D imaging Lidar system.

Sparse representation based multi-threshold segmentation for hyperspectral target detection

A sparse representation based multi-threshold segmentation (SRMTS) algorithm for target detection in hyperspectral images is proposed. Benefiting from the sparse representation, the high-dimensional spectral data can be characterized into a series of sparse feature vectors which has only a few nonzero coefficients. Through setting an appropriate threshold, the noise removed sparse spectral vectors are divided into two subspaces in the sparse domain consistent with the sample spectrum to separate the target from the background. Then a correlation and a vector 1-norm are calculated respectively in the subspaces. The sparse characteristic of the target is used to ext ract the target with a multi -threshold method. Unlike the conventional hyperspectral dimensionality reduction methods used in target detection algorithms, like Principal Components Analysis (PCA) and Maximum Noise Fraction (MNF), this algorithm maintains the spectral characteristics while removing the noise due to the sparse representation. In the experiments, an orthogonal wavelet sparse base is used to sparse the spectral information and a best contraction threshold to remove the hyperspectral image noise according to the noise estimation of the test images. Compared with co mmon algorithms, such as Adaptive Cosine Estimator (ACE), Constrained Energy Minimizat ion (CEM) and the noise removed MNF-CEM algorithm, the proposed algorithm demonstrates higher detection rates and robustness via the ROC curves.

An optimized hybrid encode based compression algorithm for hyperspectral image

Compression is a kernel procedure in hyperspectral image processing due to its massive data which will bring great difficulty in date storage and transmission. In this paper, a novel hyperspectral compression algorithm based on hybrid encoding which combines with the methods of the band optimized grouping and the wavelet transform is proposed. Given the characteristic of correlation coefficients between adjacent spectral bands, an optimized band grouping and reference frame selection method is first utilized to group bands adaptively. Then according to the band number of each group, the redundancy in the spatial and spectral domain is removed through the spatial domain entropy coding and the minimum residual based linear prediction method. Thus, embedded code streams are obtained by encoding the residual images using the improved embedded zerotree wavelet based SPIHT encode method. In the experments, hyperspectral images collected by the Airborne Visible/ Infrared Imaging Spectrometer (AVIRIS) were used to validate the performance of the proposed algorithm. The results show that the proposed approach achieves a good performance in reconstructed image quality and computation complexity.The average peak signal to noise ratio (PSNR) is increased by 0.21~0.81dB compared with other off-the-shelf algorithms under the same compression ratio.