This paper proposed a lens assisted two-dimensional beam scanning and ranging device based on the control of field programmable gate array (FPGA). A lidar system with 16 scanning angles, 19.5 m ranging distance, and 2 cm ranging error was achieved through the Mach-Zehnder interferometer (MZI) optical switches with a binary tree structure, FPGA control, and pulse width modulation (PWM). Our work shows the potential of efficient FPGA control in the lens assisted beam scanning technology. The control scheme selected in this paper only needs log N channels of PWM waveforms for N optical transmitting ports (N beam scanning angles). Compared with the typical optical phased array (OPA) technology that needs N channels of digital-to-analog converter (DAC) for N optical transmitting ports (the control complexity is O(N)), this scheme greatly reduces the control complexity (O(log N)) and system cost. Therefore, it is expected to pave the way towards the application of lidar systems based on integrated beam scanning technology.

The absorption and scattering of solar radiation by the atmosphere will reduce the brightness and contrast of satellite images. The lower the atmospheric visibility and the higher the satellite spatial resolution, the more obvious this phenomenon, so that the sub-meter spatial resolution optical satellite image under low visibility conditions looks very blurry. The adaptive atmospheric correction algorithm developed based on the radiative transfer equation fully considers the influence of the atmosphere and the surrounding environment of the target on the target radiance at the entrance pupil of the satellite, and quantitatively describes the influence of the reflectance difference between the background pixels and the target pixel on the adjacency effect. The adaptive atmospheric correction algorithm is utilized to perform atmospheric correction on sub-meter spatial resolution satellite images under low atmospheric visibility conditions, and the results are compared with processing results of conventional images. The results show that the quality of the satellite image corrected by the adaptive atmospheric correction algorithm has been significantly improved (the image sharpness increases by 4.5275 times, the image contrast increases by 44.61%, and the image information entropy value increases by 64.22%). Compared to conventional image processing methods that will bring noise and excessive enhancement when improving the quality of satellite images, the adaptive atmospheric correction algorithm will not bring noise and excessive enhancement when improving the quality of satellite images.

In this paper, we analyzed the characteristics of the mid-infrared spectra containing the 13CO2 strong absorption with a transmission model of scanning imaging absorption (SCIATRAN) and a high-resolution transmission molecular absorption database (HITRAN). We found that the 13CO2 absorption was more obvious in the mid-infrared band than in the near-infrared and thermal infrared bands; the mid-infrared band was extremely sensitive to changes in the CO2 concentration in the 5--15 km middle and upper troposphere. Atmospheric convection and exchange make the change in the CO2 concentration in the middle and upper layers closely related to the near-surface. The analysis of the spectral characteristics indicates that the main factors affecting CO2 retrieval are water vapor, temperature, and nitrous oxide. Furthermore, we retrieved the 13CO2 concentration with satellite remote sensing data and provided the necessary conditions for retrieval. Specifically, the profiles of water vapor, temperature, and nitrous oxide should respectively have the precision of 50×10 -6, above 0.03 K, and above 5×10 -9; the signal-to-noise ratio of the instruments needs to be higher than 600. Our findings provide a theoretical basis for 13CO2 detection with satellite remote sensing and put forward requirements for indicators for future hyperspectral detector manufacturing. As such, we can better control the CO2 source, sink, and transmission laws and grasp the global carbon cycle budget.

In order to analyze the influences of rough surface parameters on the statistical characteristics of laser speckles, it is necessary to study the mapping relationship between laser speckles and target materials. Random rough surfaces with different combinations of root-mean-square roughness, correlation length, skewness, and kurtosis are simulated by computer, and the laser speckle patterns generated by these random surfaces are analyzed and processed based on the theoretical model of the laser speckle field. The results show that various surface characteristic parameters will affect the statistical characteristics of the laser speckle field, and the laser speckle fields formed under different combinations of surface parameters show great specificity. Therefore, the statistical characteristics of laser speckles studied in this paper may be used as effective information to distinguish the surface of materials in the future.

When traditional diffraction lenses work in a wide band, the results have severe dispersion and poor image quality. Thus, an optimized design method for a partition diffractive achromat based on RGB three-band is proposed. We use scalar diffraction theory to simulate and analyze the image of equal width diffractive achromats and equal area diffractive achromats which compare with traditional diffraction lenses in point spread function and diffraction efficiency. The analysis results show that the proposed partition diffractive achromat can reduce the standard deviation of the three-wavelength diffraction efficiency from 0.6607 of the traditional diffractive lens to 0.1519 and 0.0592 when working in the RGB three-band while maintaining imaging effects well. Finally, considering the actual processing conditions, the optical performance parameters corresponding to the microstructure quantization into eight steps under different partitioning methods were simulated. The results show that the proposed design method can make the diffraction lens achieve a good achromatic effect in the RGB three-band. Moreover, the proposed method has practicability and universal applicability.

Owing to the lack of a particular function to fit the channel impulse response (CIR) curves in clear ocean water, a method of fitting CIR curves of clear ocean water using double exponential function is proposed in this article. The proposed double exponential function is compared with four other well-known fitting functions—single Gamma, inverse Gaussian, double Gamma, and CEAPF—to show its superiority. Moreover, we compare and analyze the fitting performance and working range of these five functions under different conditions of seawater quality, link range, and field of view (FOV) of the receiver, and these are verified by Monte Carlo simulation experiment. Both theoretical analyses and simulation results suggested that the proposed double exponential function could achieve an accurate fitting performance not only in clear ocean water but also in harbor water and could outperform the four other functions, especially in a short link range of clear ocean water (e.g., 10 m) and in harbor water case with different link ranges and FOVs. From the results, the single Gamma function used to match the CIR curves in the atmospheric environment could also be used to fit the ones in clear ocean water and harbor water, whereas the inverse Gaussian function deviated the CIRs under any working case with larger errors; the double Gamma function could fit the CIR of coastal water and harbor water but could not fit that of clear ocean water; the CEAPF function could generally fit well with the CIRs in the three ocean water qualities and obtain good fitting results.

Aiming at addressing the intercell interference in an indoor visible light communication system, aperture array receivers are designed and optimized herein. The aperture array receivers comprise a bare central detector and various receiving elements and use avalanche photodiodes as detectors. In the optimization problem, the maximum signal-to-interference-and-noise ratio is assumed as the optimization objective, and the interference-to-signal ratio is constrained to zero. The optimal structure parameters of the aperture array receivers are obtained under two receiving modes at the center of the room and ergodic positions. The results show that compared with generalized angle diversity receivers and a single avalanche photodiode, the bit error rate of the proposed receivers is decreased by at least 10 dB and 14 dB, respectively, in the center of the room. Furthermore, the signal-to-interference-and-noise ratio of the proposed receivers is improved by 6 dB and 15 dB, respectively, in ergodic indoor positions.

In this work, a method based on the traditional Gerchberg-Saxton (GS) algorithm and convolutional neural network (CNN) is proposed to generate vortex beams using a liquid crystal spatial light modulator (LC-SLM). By adopting this GS-CNN method, the Bessel beams with different topological charges are generated. On this basis, the root mean squared error (RMSE) and diffraction efficiency (DE) of the generated vortex beams are further analyzed and compared with the results obtained by the traditional GS algorithm. The results show that the GS-CNN method proposed in this paper can produce high-quality Bessel vortex beams. Compared with the results from the traditional GS algorithm, the intensity difference between the generated vortex beam and the target light is reduced and there are more input light field energies to be diffracted.

Computational ghost imaging uses the second-order coherence of light fields to reconstruct images. In the case of an unknown disturbance (like atmospheric turbulence) to the probe light, the actual light field reaching the object cannot be calculated, and the images will become blurred if they are reconstructed according to the calculated light field without disturbance. In this paper, we proposed a deep learning based image classification-restoration method to suppress the influence of atmospheric turbulence on computational ghost imaging. Specifically, the classification network based on a convolutional neural network classified images according to their blur degree. Then, the images of each class were restored by the restoration network based on a generative adversarial network. Furthermore, we established a compressive-sensing-based computational ghost imaging model including atmospheric turbulence. As a result, the blurred images caused by atmospheric turbulence of different intensities were obtained, and the blurred images were classified and restored by the deep learning method. The simulation results show that the proposed classification-restoration network can effectively improve the image quality of ghost imaging and significantly improve the structural similarity and peak signal-to-noise ratio of the restored images. Besides, this network can generalize different types of targets.

The Fresnel zone plate (FZP) enabling the focusing of light sources is one of the main components in hard X-ray microscopy. Resolution and diffraction efficiency are the two most important parameters of FZP, which, however, are often difficult to be considered at the same time in actual design and preparation. Therefore, a design method for hard X-ray FZP based on rigorous coupled wave theory is proposed in this paper. The method optimizes the diffraction efficiency based on the specified resolution and gives the optimized values of the composition materials, the width of the ring zone, the outer diameter, the thickness, and the thickness control accuracy associated with the hard X-ray FZP. Furthermore, considering the influence of material dispersion, the distribution of the optimal diffraction efficiency with the change in the light source energy is given, which provides a reference for the light source selection in microscopic imaging.

The large-scale distributed measurement system is a measurement network based on multiple measurement fusion. The network structure is the key to improve the network performance, which has an important impact on the measurement accuracy, efficiency and even cost. To solve the network structure reconstruction caused by the changing of obstacles, measurement objects and requirements, based on the coverage performance of the network nodes, the method to determine the lost-of-light issue based on the fast collision detection algorithm is first investigated and the problem of blind spot judgement is solved. Then, the network reconstruction algorithm based on Next-Best-View (NBV) is studied. At the same time, in order to improve the networking efficiency, the improved grey wolf optimization algorithm is used as the best location search algorithm to rearrange the numbers and positions of nodes, and thus efficient networking is realized and the reconstruction accuracy is improved. Finally, with the workshop Measurement Positioning System (wMPS) as the verification platform and from three aspects of measurement coverage, accuracy and networking efficiency, the effectiveness of the proposed method is confirmed by changing measurement conditions and requirements.

The surface quality of precision metal parts will affect the performance and appearance of the product. For this reason, it is necessary to measure and evaluate the surface texture of the parts. Commonly used contact measurement methods have low measurement efficiency, and optical measurement methods are susceptible to the effects of surface highlight reflection. For this reason, a method based on near-field non-Lambertian photometric stereo vision for highlight metal surface texture reconstruction is proposed. In order to effectively describe the non-Lambertian reflection, this method uses an inverse reflectance model based on an inverse reflectance light source to decouple the surface normal vector and the nonlinear reflection model. This method uses neighborhood information to improve the robustness of the inverse model, and uses the maximum fusion strategy to overcome the influence of shadows and rendering generates targeted simulation datasets, thereby improving the adaptability to metal surface reflection. The results show that the proposed method can reconstruct the high-brightness metal surface texture with high precision, and the relative measurement error is less than 15%.

After the parameter calibration of a projector, it can be considered another camera and combined with a binocular system to form a trinocular stereo vision system. In contrast to the traditional calibration method using high-precision plane targets, the proposed method only needs a simple and unmarked white paper target to obtain the world coordinates of the characteristic points provided by a binocular system. The linear model and distortion correction model containing the distortion coefficient of a thin prism are used to complete the preliminary parameter calibration of a projector, and an improved beam adjustment method is used to estimate the global optimal parameters of the projector and a dual camera. Results of calibration verification experiments show that the proposed improved beam-adjustment method can achieve parameter optimization. A measurement comparison experiment on a standard plane showed that the resulting trinocular stereo vision system could enhance measurement accuracy.

It is an effective method to improve the automation level of aircraft skin damage detection by using machine vision. The global three-dimensional (3D) reconstruction of aircraft skins is the key step in the detection process to locate the exact location of the damaged parts. In order to solve the problems of complex equipment, low splicing accuracy, and low processing efficiency in current technology, this paper proposes a rapid 3D reconstruction method for aircraft skins based on rear positioning and combined with structured light. The 3D reconstruction of the local area of aircraft skins is conducted by using the structured light 3D measurement system, and the rear camera synchronously observes the structured light system to determine its spatial position and posture. With the help of spatial pose data, the 3D topography data of the aircraft measured by the structured light system is fused into the positioning camera coordinate system to realize the rapid and high-precision non-contact measurement of the 3D topography of the large aircraft skin. This method provides effective technical support for the automatic visual inspection of aircraft skin damage.

High quality-factor, low-loss and high-transmission electromagnetically induced transparency is one of the important research directions of metamaterials. In this paper, a cross-shaped all-dielectric metamaterial structure is designed, which is composed of two mutually orthogonal silicon empty cavities. When the cross-shaped structure is fully symmetrical, there is no electromagnetically induced transparency. However, when the symmetry of the structure is broken in a specific direction, the dark mode is introduced and the electromagnetically induced transparency can be observed. The transmission peaks with 93% transmissivity and 1064 quality factor can be realized simultaneously. The physical mechanism of the electromagnetically induced transparency is analyzed using the electromagnetic resonance modes, the influences of asymmetric offsets under different directions, geometric structrue and refractive index on the transmission spectrum are further investigated, and the high group refractive index phenomenon at transmission peaks is analyzed using phase change. These characteristics make the proposed metamaterial structure possible to be applied in slow-light devices, sensors and optical switches.

Photonic crystals, artificial microstructures composed of periodic dielectric materials, have broad application prospects in integrated optoelectronics, nanophotonics, and other fields. In this paper, the fabrication of compound photonic crystals by the multi-beam holographic interferometry is studied theoretically and numerically. According to the multi-beam interference principle, we design a beam configuration of compound photonic crystals and write a program by MATLAB. The simulation results agree well with the theoretical predictions. Furthermore, we symmetrically study the influences of single-beam, two-beam, three-beam, and four-beam polarization combinations on the unit cell of the compound photonic crystals. The results indicate that polarization combinations have a significant influence on the unit cell, and various unit-cell shapes such as double-drop and double-circle shapes can be easily obtained at different polarization combinations. The optimal contrast is achieved when all beams are linear polarization. Additionally, the influence of the initial phase of the incident light on the unit cell is discussed. The above research guides the design and fabrication of compound photonic crystals with various unit-cell shapes and can be used in the physics experiment teaching.

Due to the special layered structure and appropriate band gap, BiOX (X= Cl, Br, I) shows good photocatalytic activity and stability in visible light. Changing the composition of halogen in BiOXs can adjust the band gap, thereby optimizing the photocatalytic performance. In this paper, a series of BiOClxBryIz(x, y, z=0.1, 0.2, 0.33, 0.6, or 0.8) composite catalysts were prepared by a mixed solvothermal method, and their structural morphologies and optical properties were characterized. Furthermore, the photocatalytic performance of these catalysts was investigated by degrading 15 mg/L methyl orange (MO) under a 300 W xenon lamp with a 400 nm cutoff filter for 180 min. The results show that BiOClxBryIz composite catalysts displayed a microsphere structure assembled by nanosheets and their degradation efficiencies towards MO were higher than those of pure BiOX (X= Cl, Br, I) except BiOCl0.8Br0.1I0.1. Among the prepared BiOClxBryIz composite catalysts, BiOCl0.33Br0.33I0.33 composite catalyst exhibits the best degradation performance, its degradation efficiency can reach 98.4%. The reason may be that the composite catalyst with a more suitable band gap can make better use of visible light and improve the separation and transmission efficiency of photogenerated carriers.

The Li2MoO4 powder was synthesized and its crystal was grown at room temperature by the aqueous solution method, respectively. Compared with the common high-temperature sintering, this method effectively improved the uniformity and purity of the powder. The Li2MoO4 ingot has a regular shape with a size of 20 mm×30 mm×40 mm. The crystal grows regularly in the crystal-face directions and its main crystal planes are (100) and (110) planes. The morphology of Li2MoO4 was observed by a metallographic microscope and the crystal was of step-like growth. The cutoff wavelength on the absorption edge of the crystal is about 292 nm. At room temperature, the wavelengths of excitation and emission peaks are 370 nm and 445 nm respectively, while they shift to 275 nm and 520 nm respectively at low temperatures (10 K). As the temperature decreases, the wavelength of the emission peak has a redshift and the emission intensity increases sharply.

During the inversion of optical parameters of biological tissues, the measurement accuracy is low and the in vivo measurement is difficult. Therefore, a neural network model to invert the optical parameters of biological tissues was proposed in this paper. In this method, the diffuse reflectance R(r) at different detection distances r from the Monte Carlo algorithm is used as the input, and the absorption coefficient and scattering coefficient are taken as the output. The absorption coefficient and scattering coefficient retrieved by the neural network algorithm are compared with those by the Monte Carlo algorithm. The simulation results show that with the diffuse reflectance at r=0.1 cm and r=0.3 cm as the input, the mean absolute errors are 0.003 and 1.574, respectively for the absorption coefficient and scattering coefficient retrieved by the neural network algorithm, and the consistency coefficient of determination R2 can reach 0.9997 and 0.9915, respectively. The biological tissue parameters retrieved by the neural network model agree well with the absorption coefficient and scattering coefficient obtained by the Monte Carlo algorithm. The neural network model has the advantages of high inversion accuracy and simple operation, which provides a new method for the in vivo measurement of optical parameters of biological tissues.

The spectroscopic imaging system in a mid-wave infrared imaging spectrometer was designed and developed for hyperspectral remote sensing satellites on geostationary orbit. According to the characteristics of geostationary orbit, index parameters of the optical system were analyzed. It was pointed that such spectroscopic imaging system features long slit, small relative aperture, and compact size. A double-slit Offner spectroscopic imaging system was designed with a spectral range of 3--5 μm, splicing slit length of 49 mm, spectral sampling distance of 50 nm, and F-number of 5.4. The whole machine refrigeration was adopted to improve optical system sensitivity. The analysis results showed that when the working temperature was lower than 160 K, the noise equivalent temperature difference (NETD) reduced to 0.2 K, and an athermal design was performed on the basis of the cooling temperature. Based on the results of optical design, the whole aluminum alloy imaging system prototype was developed. The peak efficiency of a self-developed convex blazed diffraction grating reached 86.4%. The performance of the prototype at room temperature was tested using high order diffraction spectrum of visible and near-infrared light. The test results showed high imaging quality of the imaging spectrometer with slight smile and keystone distortions; in addition, all specifications met the design requirements.

In order to achieve a high degree of light weight, a sector-shaped SiC light-weight primary mirror of a 1000 mm photoelectric theodolite is taken as the research object and optimized design. By optimizing the thickness of the ribs on the back of the main mirror, the thickness of the semi-closed back panel and the total thickness of the main mirror, the volume and quality of the main mirror are reduced. After the three-dimensional model is established, the finite element model is established using the finite element software Abaqus, and the deformation analysis of the primary mirror is performed after the finite element simulation results are obtained. The Zernike polynomial is used to fit the deformation data of the primary mirror, and the root mean square (RMS) value of the surface error of the primary mirror is obtained. The simulation results show that the mass of the optimized main mirror is 62.78 kg, which is 30% lower than the initial mass (89.36 kg), and the diameter-to-thickness ratio of the main mirror is increased from 8.58 to 11.44 when the design requirements of the main mirror surface accuracy are met. When the optical axis is horizontal, a four-dimensional interferometer is used to detect the surface shape of the fan-shaped lightweight main mirror, and the detection result of the RMS value of the main mirror surface error is 18.22 nm.

A new type of light-controlled dual-directional gate silicon controlled rectifier (LDGSCR) protection device is designed by using the photovoltaic effect, and the regulation and control effect of light on the maintenance window of electrostatic discharge (ESD) is studied. The photo-generated current is used to simulate the physical effects of the light control ESD design window, and a macroscopic model of the silicon controlled rectifier (SCR) of the light control device is realized. Under the simulated illumination of 1.5 V voltage, the holding current of this model is increased by 35 mA compared to the no-light condition, indicating that the use of light to adjust the ESD maintenance window can reduce the risk of latch-up of the protected circuit. The LDGSCR device is fabricated and tested using a 0.18 μm BCD process. The maximum error between the test result and the model simulation result is only 0.09 V and 0.004 A. The verification shows that the macro model can eliminate the convergence problems existing in the traditional coupled transistor circuit model, and greatly reduce the time and effort required to develop a new structure of light-controlled SCR devices.

In order to study the influence of black carbon aerosol on the performance and transmission quality of satellite-ground quantum communication, the extinction efficiency factor of black carbon aerosol is obtained based on its complex refractive index and the Mie scattering theory. In addition, the influence of black carbon aerosol particle number concentration on the link attenuation, channel capacity, channel survival function and channel error rate of quantum communication is analyzed and simulated. The results show that when the transmission distance is 2 km and the black carbon aerosol particle number concentration increases from 1.2×10 7 m -3 to 3.6×10 7 m -3, the link attenuation increases from 0.5079 dB to 1.524 dB, and the bit error rate increases from 0.006 to 0.0101. Simultaneously, the channel capacities and channel survival functions of the amplitude damping channel and the depolarization channel both show different degrees of attenuation, in which those of amplitude damping channel attenuate more obviously. Therefore, black carbon aerosol has a significant impact on the performance of satellite-ground quantum link communication. In order to ensure the smooth progress of quantum communication and improve the reliability of transmission system, the relevant strategies and parameters should be adjusted to reduce the impact of black carbon aerosol on quantum communication links.

With the advent of the 5th generation (5G) wireless systems, the development of mobile communication networks with large capacity, high frequency, and multi-service access is confronted with enormous challenges. In the 5G network, since the high-frequency signals are adopted, the coverage of a base station is greatly reduced and the number of base stations to be deployed is dramatically increased. Thus, the base stations tend to be miniaturized, densified, and passive for a large total coverage range together with lower overall costs and power consumption. The power-over-fiber (PoF) technique applied to radio-over-fiber (RoF) systems provides a solution to a new generation of distributed micro base stations. This technique, by transmitting the radio-frequency signals and power light to the base stations via optical fibers, achieves the power supply for the base stations and the emission of the wireless broadband signals with antennas. In this paper, we introduced the recent research progress of the PoF for RoF systems and analyzed its development trend by summarizing the features of typical schemes.

To find alloy nanoshells with better light absorption and backscattering properties, we quantitatively analyze the effects of the core radius, shell thickness, alloy composition, and ambient medium of Au-Ag alloy nanoshells on the light absorption and backscattering properties. In the process, the Mie scattering theory of double-layer concentric spheres and the size correction model of dielectric functions are adopted. The results show that when the Au molar fraction of 50%, the size step of 0.01 nm, the inner core is SiO2 and vacuum, the maximum volume absorption coefficients of Au-Ag alloy nanoshells are respectively 93.660 μm -1 and 99.316 μm -1, with the inner core radius of 27.89 nm and 28.02 nm and the shell thicknesses of 3.95 nm and 3.35 nm. When the maximum volume backscattering coefficients are respectively 5.280 μm -1 and 5.550 μm -1, with the inner core radius of 56.08 nm and 56.37 nm and the shell thicknesses of 10.47 nm and 8.89 nm. When the Au molar fraction is less than 9%, the light absorption characteristics of Au-Ag alloy nanoshells are better than those of Au nanoshells. When the Au molar fraction is lower than 11%, the Au-Ag alloy nanoshells are superior to Au nanoshells in backscattering characteristics.

To study the polarization characteristics of metal surfaces, we analyze the physical mechanism of the interaction between light and object. First, the reflection on a target surface is divided into specular reflection, diffuse reflection, and volume scattering according to the scattering characteristics of microfacets, through which a three-component polarization bidirectional reflection distribution function model is built in this paper. Then, the model parameters are optimized by the nonlinear least square method. The simulation results based on open-source data show that in comparison with other functions, the curves fitted by the function are more consistent with the measured results and the polarization characteristics of different metal surfaces can be exactly described; moreover, the reflection information on the surface of metal materials can be accurately disclosed. This study provides new ideas for investigating the polarization properties of metal surfaces.

The current ammonium perchlorate water content detection methods are difficult to meet the rapid and safe detection requirements. The time domain spectra of samples were first measured by the terahertz time domain spectroscopy technique and then the refractive index and the absorption coefficient were extracted. According to the Lambert-Beer law and the effective medium theory, the water content prediction model based on absorption coefficients was constructed. Simultaneously, according to the effective medium theory and the equal optical paths, the water content prediction model based on refractive indexes was constructed. These two prediction models both need to weigh the samples. In addition, this paper proposed another new water content prediction model based on absorption coefficients combined with refractive indexes, which does not need to weigh the samples under the condition of ensuring the measurement accuracy. The experimental results showed that when the water mass fractions of the samples in the same group were below 1%, the determination coefficients between water content and absorption coefficient or refractive index were both greater than 0.995. In contrast, when their water mass fractions were below 0.1%, the determination coefficients were around 0.95. The root mean square errors of prediction for these three models were all within 0.05%. The baseline shift between the predicted and measured water contents for different groups of samples is probably caused by the difference in sample preparation. The study shows that the proposed method provides a rapid, nondestructive, safe and high precision detection method for the detection of water contents in ammonium perchlorate and similar solid powder.

Digital cameras usually build a colorimetric characterization model through target samples, which, however, will affect the mapping between the cameras’ response values and the colorimetric values. The existing selection methods for target samples ignore the generality of target samples, namely that the colorimetric characterization model built based on the optimized target samples should be applicable to any color samples. Therefore, we proposed an optimization method of colorimetric characterization targets for the digital cameras based on uniform color samples. First, uniform color samples were selected from the spectral reflectance of extensive real objects. Then, the colorimetric characterization model was combined to optimize the target samples among the uniform color samples. The results show that the optimization of target samples is subject to the data type and the colorimetric characterization model. Consequently, different optimized target samples are obtained, and the optimized target samples outperform the typical target samples in terms of colorimetric characterization. In addition, for the RAW data, an increase in the number of target samples does not necessarily improve the accuracy of a linear colorimetric characterization model, while for the sRGB data, more target samples will be beneficial for enhancing the performance of a polynomial colorimetric characterization model.

Color constancy is an important prerequisite for visual tasks such as recognition, segmentation, and three-dimensional object reconstruction. We proposed a multi-channel feature-confidence-weighted network to enable the computer vision systems to perceive color constancy. As a result, the network could fully extract the features in the images while reducing the number of network layers and model parameters. The multi-channel confidence-weighted method employed the features in each channel that could provide more information for light source estimation to accurately estimate the light source in the global scene. Experimental results on the reprocessed ColorChecker and NUS-8 datasets show that the proposed algorithm, which weights the confidence of features in multiple channels, outperforms its counterparts in terms of all evaluation indexes and thus has higher accuracy and robustness. As such, this algorithm can be applied to the tasks of computer vision requiring color correction.