Overview: Numerous sub-aperture fiber laser array is one of the emerging technologies to build high power, high beam quality and equivalent optical large aperture. Realizing the common phase and even the fast and flexible beam deflection of array laser beam based on the precise phase control is the key to the application of the current fiber laser phased array technology. In this paper, the optical phase-controlled steering technology is combined with the fiber laser coherent combining system, and the beam steering characteristics of the numerous sub-aperture, meter-scale fiber array laser coherent combining system are studied. Aiming at the development trend of numerous sub-aperture fiber laser phased array technology, based on the 19 aperture fiber laser phased array as the basic module, the meter-scale phased array transmitting system models with 19, 133 and 703 apertures are established. Based on the principle of optical phased array, the step phase folding model is adopted to make the piston phase distribution of the beam emitted from adjacent aperture change continuously, and to realize the high-precision continuous steering in a certain range. Meanwhile, the steering limit ranges of 19, 133 and 703 aperture fiber laser phased arrays are defined and calculated according to the distribution characteristics of the far-field steering beam pattern. Through numerical simulation analysis, the results show that when the piston phase difference of adjacent sub-apertures changes at equal intervals, the far-field main lobe position changes, and the steering angle gradually increases with the increase of phase difference. When the steering angle increases, the far-field main lobe energy gradually leaks into the grating lobes, which reduces the peak light intensity of the main lobe. When the peak intensity of the grating lobe is stronger than the main flap, the energy concentration of the steering beam on the far-field target surface is poor, which easily affects the position calculation of the far-field main lobe and interferes with the precise pointing control of the steering beam. Therefore, the limit range of steering is defined when the peak intensity ratio of the main lobe to the grating lobe is equal to 1. When the fiber laser phased array steers along the x- and y-axes respectively, there are obvious differences in the far-field spot shape and steering range, which is caused by the asymmetric structure of the fiber laser phased model. In this paper, the phased array models with apertures 19, 133 and 703 have equivalent diameters. As the number of sub-aperture increases, the aperture spacing decreases and the steering range increases. Therefore, the parameters of the phased array steering system can be designed according to the actual application scenario, and the aperture size and aperture number can be selected reasonably. By studying the steering characteristics of numerous sub-aperture and meter aperture fiber laser phased arrays, this paper enriches the beam wavefront control ability of fiber laser phased array technology, which can be used for precise tracking of ultra-long-distance targets and fast beam coverage in a certain range.Numerous sub-aperture fiber laser array is one of the emerging technologies to build high power, high beam quality and equivalent optical large aperture. Realizing the common phase and even the fast and flexible beam deflection of array laser beam based on the precise phase control is the key to the application of the current fiber laser phased array technology. In this paper, the optical phase-controlled steering technology is combined with the fiber laser coherent combining system, and the beam steering characteristics of the numerous sub-aperture fiber array laser coherent combining systems are studied. The beam steering is realized by changing the phase difference between the adjacent sub-aperture of the collimated laser array. The far-field steering beam pattern distribution characteristics of 19, 133 and 703 aperture fiber laser phased arrays are compared and analyzed, and the steering limit range is defined and calculated accordingly. The results provide a theoretical basis for the subsequent experimental research on the precise pointing control of fiber laser phased arrays under long-range transmission.

Overview: With the development of computer vision, people increasingly need to understand images, including recognizing the scenes and the human behaviors in images. The task of HOI detection is to locate humans and objects in images and infer their relationships. This requires not only locating a single object instance, but also identifying the interaction between the objects. However, machines cannot know which object humans are interacts in. Most of the existing methods solve this problem by completely pairing the people and objects. They use off-the-shelf object detectors to detect instances, but this does not meet the requirements of the HOI task. This paper proposes an object detector suitable for HOI detection based on relational reasoning, which makes use of the interactive relationship between humans and objects in the images to recommend human-object pairs, so as to reduce the occurrence of non-interactive human-object pairs as much as possible. Our method follows the two-stage detection like most works. Firstly, the interactive instance proposal network (IIPN) is used to recommend human-object pairs. The IIPN follows the pipeline of faster RCNN, but replaces the region proposal network (RPN) with the IIPN. The IIPN selects human-object pairs based on the interaction possibility between humans and objects using the visual information in the picture. It passes the message through the iterative reasoning of the graph neural networks (GNNS), only human-object pairs that include interactive relationships are selected as the IIPN’s outputs. Secondly, we design a cross-modal information fusion module (CIFM), which calculates the fusion attention according to the influence of different features on the detection results, and performs weighted fusion. This is because the existing methods simply add or splice several features such as human visual features, object visual features, and human-object spatial features in the reasoning part. The different influence degrees of various features in different actions are ignored. For example, the verbs like ride and hold in and depend more on the spatial relationships, while eat and cut in and depend more on human's postures, that is, visual features. Meanwhile, this paper believes that semantic prior knowledge is also helpful to HOI detection. For example, if we have apples in an image, the probability of predicting the human's action as eating or holding is greater than others. Finally, complete experiments are performed on two popular large-scale HOI datasets, HICO-DET and V-COCO. The experimental results show the effectiveness of the proposed method.Human-object interaction detection is to locate and identify the interactive relationship between humans and objects in an image. The challenge is that the machine cannot know which object the person is interacting in. Most existing methods try to solve this problem by matching humans and objects exactly. Different from them, this paper proposes an interactive instance proposal network based on relational reasoning to adapt to the task. Our main idea is to recommend human-object pairs by using the potential interaction relationships in the visual relationship between humans and objects. In addition, a cross-modal information fusion module is designed to fuse different context information according to its influence on the detection result, so as to improve the detection accuracy. To evaluate the proposed method, we performed sufficient experiments on two large-scale datasets: HICO-DET and V-COCO. Results show that our method achieves 19.90% and 50.3% mAP on HICO-DET and V-COCO, which are 4.5% and 2.8% higher than our baseline, respectively.

Although the FMCW LiDAR measurement technology has gradually matured, further exploration and research is still needed. At present, most of the nonlinear correction methods focus on the optical system and signal processing, but we hope to solve the problem from the design of the laser itself, and make further improvements in its mechanical structure, circuit design, and temperature control to avoid subsequent complicated work. In order to achieve true intelligence, on the one hand, we need to improve the efficiency of measurement (especially three-dimensional imaging). On the other hand, we must strive to miniaturize and integrate the FMCW LiDAR measurement system to bring more convenience and wider application scenarios.In modern measurement technology, frequency modulation continuous wave LiDAR combines the advantages of traditional radar and laser interferometry and plays an important role in the fields of the large-size space precision measurement, micro-distance measurement, and three-dimensional imaging with its characteristics such as non-contact, large measurement range, high resolution, and strong anti-jamming capability. However, in practical application, the frequency modulation of the laser light source can’t be completely linear, which greatly reduces the measurement accuracy of the frequency modulation continuous wave LiDAR technology. Therefore, how to suppress the effects of the laser frequency modulation nonlinearity has become a hot research topic in the field of frequency modulation continuous wave LiDAR measurement. This paper introduces the basic principle of the frequency modulation continuous wave LiDAR, and introduces four widely used nonlinear correction methods and some special nonlinear correction methods according to the different nonlinear correction schemes of the frequency modulation, and makes summaries and prospects.

Overview: In the atmospheric environment, there are many fine particles in the air, which will lead to the absorption or refraction of light and affect the normal radiation of light. In this case, the color, contrast, saturation and detail of the image captured by the camera are often seriously affected. At present, computer vision needs to realize many high-level tasks such as pedestrian recognition, automatic driving, air navigation, remote sensing and telemetry, and these high-level tasks have a high demand for image quality. Therefore, it is of great significance to carry out single image defogging to obtain higher quality images before performing high-level tasks. In recent years, single image defogging using generative adversarial networks(GAN) has become a hot research aspect. However, the traditional GAN algorithms rely on annotated datasets, which is easy to cause over-fitting of ground truth, and usually performs not well on natural images. To solve this problem, this paper designed a GAN network incorporating dark channel prior loss to defogging single image. This prior loss can influence the model prediction results in network training and correct the sparsity and skewness of the dark channel feature map. At the same time, it can definitely improve the actual defogging effect and prevent the model from over-fitting problem. In addition, this paper introduced a new method to obtain dark channel feature map, which compresses pixel values instead of minimum filtering. This method does not need to set fixed scale to extract dark channel feature map, and has good adaptability to images with different resolutions. Moreover, the implementation function of this method is a convex function, which is conducive to embedded network training and enhances the overall robustness of the algorithm. The proposed algorithm is quantitatively analyzed in the comprehensive test set SOTS and the mixed subjective test set HSTS. The peak signal-to-noise ratio (PSNR), structural similarity SSIM and BCEA Metrics are used as the final evaluation indexes. The final result shows that our algorithm can raise PSNR up to 25.35 and raise SSIM up to 0.96 on HSTS test sets. While it comes to SOTS test sets, our method achieves the result of 24.44 PSNR and 0.89 SSIM. When we use BCEA metrics to evaluate our algorithm, we achieve the result of 0.8010 e,1.6672 r and 0.0123 p. In summary, Experimental results show that the proposed algorithm performs well on real images and synthetic test sets compared with other advanced algorithms.Single image defogging using generative adversarial networks (GAN) relies on annotated datasets, which is easy to cause over-fitting of ground truth, and usually performs not well on natural images. To solve this problem, this paper designed a GAN network incorporating dark channel prior loss to defogging single image. This prior loss can influence the model prediction results in network training and correct the sparsity and skewness of the dark channel feature map. At the same time, it can definitely improve the actual defogging effect and prevent the model from over-fitting problem. In addition, in order to solve the problem that the extraction method of traditional dark channel feature has non-convex function and is difficult to be embedded into network training, this paper introduces a new extraction strategy which compresses pixel values instead of minimum filtering. The implementation function of this strategy is a convex function, which is conducive to embedded network training and enhances the overall robustness of the algorithm. Moreover, this strategy does not need to set a fixed scale to extract the dark channel feature map, and has good adaptability to images with different resolutions. Experimental results show that the proposed algorithm performs better on real images and synthetic test-sets like SOTS when compared with other sota algorithms.

Overview: As a super-resolution optical imaging technology, structured light illumination technology carries an object’s high-frequency information into the optical system in the form of moiré fringes through structured illumination, breaking the diffraction limit in traditional optical imaging and improving image resolution. An incoherent self-interference digital holography based on the Michelson interferometer can accurately record an object's phase and intensity information. It has the characteristics of fast real-time, non-contact, non-marking, three-dimensional imaging, etc., and has essential research significance in biomedical imaging and materials science. In this paper, an incoherent digital holographic imaging system based on the Michelson interferometer with structured light illumination is proposed, which uses a spatial light modulator (SLM) to realize horizontal and vertical cosine grating illumination patterns to improve the lateral resolution of the imaging system. Perform simulation and verification experiments in uniform and structured light illumination mode to explore the high-resolution imaging results of the resolution target. We obtained in simulation imagings: First, the resolved minimum element of the resolution target is Group 4 element 3 (20.16 lp/mm) in Figure 3(e) under uniform light illumination. Then, the algorithm is used to modulate the resolution target to realize the structured light illumination mode. The resolved minimum resolution element of the resolution target is Group 5 element 2 (35.92 lp/mm) in Figure 4(c). We get in the verification experiments: First, use the algorithm to generate a mask with a value of 1 on the SLM to adjust the illumination mode to the uniform light illumination mode, and the resolved minimum resolution element of the resolution target is the Group elements 4 (45.25 lp/mm) in Figure 5(e). Using another algorithm to load cosine gratings of 20 lp/mm and 40 lp/mm on the SLM to adjust the illumination mode to structured light illumination mode, the resolved minimum element of the resolution target is Group 6 element 1 (64 lp/mm) and Group 6 element 4 (90.51 lp/mm) in Figure 6(a1) and Figure 6(b1). The applicability of the super-resolution imaging method based on the structured light illumination to the incoherent light self-interference digital holographic imaging system based on the Michelson interferometer is verified from the level of simulation imaging and experiments, and the resolution of the imaging system is improved. In the future, it is necessary to comprehensively consider the system performance, optimize the system structure, study more effective numerical algorithms, and realize super-resolution imaging, dynamic imaging, color imaging, etc., to obtain more excellent development space.An incoherent digital holographic imaging system based on the Michelson interferometer with structured light illumination is proposed, which uses a spatial light modulator (SLM) to realize horizontal and vertical cosine grating illumination patterns to improve the lateral resolution of the imaging system. Using MATLAB software to carry out simulation imaging and numerical reconstruction, the high-resolution reconstructed image under the system is obtained. It theoretically proves that this method can effectively improve the resolution of the incoherent digital holography system. And build the corresponding incoherent light self-interference digital holographic imaging system. By imaging the USAF1951 resolution target, further verified the applicability of the super-resolution imaging method based on structured light illumination experimentally.

Overview: Terahertz detector is an important device in the field of terahertz technology, and it is important to improve its sensitivity. The sensitivity of the detector can be improved in two aspects: one is to further optimize the antenna of the detector, and the other is to optimize the size of the detector and the spot size of incident terahertz wave. Due to the long wavelength of the electromagnetic wave in the terahertz band, the spot size is much larger than the effective acceptance area of the detector, which limits the effective absorption rate of the detector to the incident terahertz wave. In order to make the focus spot to be small, the lens aperture needs to be increased. At present, the commonly used method is to integrate the hyper-hemispheric silicon lens with the terahertz detector to reduce the spot size by one order of magnitude and increase the electric field energy density. However, hyper-hemispheric silicon lens is difficult to be ultra-thin and ultra-light, and is not planar, which is not conducive to the device integration, especially for array detectors. In this paper, a series of metasurface lenses for terahertz detectors are designed using sub-wavelength silicon cylinders. By tuning the diameter of the silicon cylinders, the transmission phase of the terahertz wave can be controlled from 0 to 2π with high transmission amplitude. At 1 THz, the backside integration of the designed single-surface lens with the terahertz detector can increase the electric field energy density in the core region of the THz detector to 32 times that of the incident plane wave, and reduce the focal spot to the same order of magnitude as the wavelength. Based on the feasibility of fabrication and anti-reflection considerations, we propose a two-sided metasurface lens, which further increases the energy density of the electric field to 44 times that of the incident plane wave. Compared with the traditional hyper-hemispheric silicon lenses, the size and thickness of the metasurface lens are smaller and more convenient for integration. Metasurface lenses have a great prospect for reducing the complexity of the terahertz system and improving the responsiveness of the detector, and provide a new idea for the integration and miniaturization of the terahertz device. However, the current metasurface lenses produce many side lobes after focusing, resulting in low focusing efficiency. Further research needs to further optimize the materials and unit structures of the metasurface lenses, to improve the focusing efficiency and electric field energy density.In this paper, a focusing lens for terahertz detection is designed using a metasurface composed of sub-wavelength silicon cylinders. By tuning the diameter of the silicon cylinder, the transmission phase of the THz wave is controlled from 0 to 2π. At 1 THz, the terahertz electric field energy density focused by the single-sided metasurface lens designed can be increased to 32 times that of the incident wave. After adding the anti-reflection, a double-sided metasurface lens is proposed, which is feasible in processing, increasing the electric field energy density to 44 times that of the original. Compared with the traditional hyper-hemispheric terahertz silicon lenses, our metasurface lens has the advantages of thin thickness and small volume, which is conducive to the miniaturization of the terahertz detector component and provides the possibility to realize the integration with the terahertz detector.