Rapid quantification of hazardous trace metals in waste oil remains challenging because of the high viscosity and severe matrix effects of oily substrates. Here, we developed a quantum dot-enhanced double-pulse laser-induced breakdown spectroscopy (DP-QDELIBS) platform for ultrasensitive and matrix-tolerant metal analysis in waste oil. AgNC@AgAux core-shell quantum dots were used as an interfacial enhancement layer to improve laser energy coupling and initial plasma formation, and a second laser pulse was introduced to reheat the primary plasma and further amplify metal emission. Compared with conventional single-pulse LIBS, DP-QDELIBS achieved signal enhancement factors of up to 47 for Mg emission lines. Ablation morphology analysis indicated that the enhancement was associated with porous microstructures generated by intensified interfacial charge transfer and localized plasma expansion. Under optimized conditions, the method showed good linearity for representative metals, with coefficients of determination above 0.97 and limits of detection of 0.02 ppm for Mg, 0.06 ppm for Ca, 0.05 ppm for Cr, and 0.05 ppm for Ba. Analysis of real industrial waste-oil samples, including engine oils and brake fluids, showed strong agreement with inductively coupled plasma optical emission spectrometry (R2 > 0.97). These results demonstrate the potential of DP-QDELIBS for rapid screening of hazardous metals in complex oily wastes and for waste-oil risk assessment and management.

Laser-induced breakdown spectroscopy (LIBS) is widely used in various fields for its powerful elemental analysis ability. However, the remote detection has always been a key limiting factor. As distance increases, the LIBS spectra intensity decays rapidly, while noise and interference intensify, which will significantly degrade the precision of remote LIBS analysis. To tackle this problem, we designed a remote LIBS system, which can simultaneously acquire the LIBS and laser echo data, and adopted the data fusion strategy to improve the classification accuracy of remote samples. We developed a Kolmogorov-Arnold network (KAN) classification model for samples classification, where the fused LIBS and laser echo data are used as the model input. Then we carried out the classification experiment on seven aluminum alloy samples at a distance of 50 m. The results show that compared with using only LIBS data (90.48%) and laser echo data (64.29%), our proposed data fusion strategy combined with deep learning method can raise the classification accuracy to 96.83%, and the stability and robustness of the model can be also improved. Moreover, compared with traditional models, such as LDA, PLS-DA, SVM, and CNN, the classification accuracy is increased by 15.48%, 11.51%, 7.54%, and 5.16% respectively. Our method offers an effective way for high-precision classification of remote LIBS, and it expands the application scope of LIBS technology, making it more promising in remote target recognition, online metal material monitoring, and other fields.

Objective The foundation of full-waveform light detection and ranging (FW-LiDAR) for acquiring target's information lies in establishing the relationship between target's features and echo waveforms, and accurately characterizing the laser echo waveforms modulated by the target. To construct the relationship between target's features and echo waveforms, researchers have proposed analytical methods, which use mathematical analytical formulas to establish a direct mapping relationship and achieve high-precision inversion of the target's lateral structural details under vertical illumination conditions. However, in practical scenarios, targets are rarely perpendicular to the illuminating laser beam. To adapt to the general situation of sloped targets, it is necessary to accurately identify the target's shape and invert multi-dimensional characteristic parameters. This paper proposes a convolutional neural network (CNN) shape discrimination algorithm that fuses attention mechanisms (AM) and feature engineering (FE), realizing intelligent discrimination of the shapes for typical two-dimensional (2D) sloped targets, and selecting the corresponding echo waveform analytical formulas to invert multi-dimensional characteristic parameters such as target size and inclination angle. Methods The shape discrimination algorithm of FE-AM-CNN proposed in this paper is shown in Fig. 6. First, the echo waveforms of the target at different positions in four motion directions are input into the system after smooth preprocessing. Key features are extracted through feature engineering and spliced with the waveform sequence. Then a pseudo-image dataset is constructed based on the target echo database. Subsequently, the dataset is used to train the AM-CNN, and the trained network is used to realize the shape discrimination of 2D sloped targets. Based on shape discrimination results, this paper selects the echo waveform analytical formula of the identified shapes. The analytical formulas of echo waveforms corresponding to different positions in the target's motion direction are integrated into a system of equations, as shown in Eq. (19). Peak intensity, peak time, and motion step width serve as known observational inputs. The unknowns in the equation system are solved via nonlinear fitting, enabling joint inversion of target size, inclination angle, and initial position. Results and Discussions Through the analysis and processing of target echo waveforms in the database, the effectiveness of the proposed intelligent target shape discrimination and multi-dimensional characteristic parameter inversion method is verified. The shape discrimination results of typical 2D sloped targets show that the discrimination accuracy of this method reaches 95.77 degrees o, 89.99 degrees o, 88.83 degrees o, and 82.87 degrees o under noise intensities ofp = 0.01, 0.1, 0.25, and 0.5, respectively (Table 5), demonstrating good noise robustness. Under the noise intensity of p = 0.25, when the target area is no less than 10.19 degrees o of the laser spot area, the algorithm can complete the discrimination of shapes except squares and circles; when the target area is no less than 40.74 degrees o of the laser spot area, all shapes can be effectively discriminated. The multi-dimensional characteristic parameter inversion results show that when the target is much smaller than the laser spot area (only 2.55 degrees o of the spot area), the inversion size error and inclination angle error are 13.31 degrees o and 6.42 degrees , respectively [Fig. 11(a)], demonstrating excellent high-resolution detection capability for small targets at long distances. Meanwhile, the characteristic parameter method proposed in this paper has strong expandability. By increasing the number of unknowns, it can enhance the dimension of information acquisition and invert characteristic parameters such as the target's initial position and motion direction. Conclusions This paper focuses on the shape discrimination of typical 2D targets and the inversion of characteristic parameters, proposing a shape discrimination algorithm of FE-AM-CNN that fuses attention mechanisms and feature engineering, and expanding the application scope of the multi-dimensional characteristic parameter inversion method. It achieves a shape discrimination accuracy of 82.87 degrees o under high noise intensity condition, and realizes the inversion size error of 13.31 degrees o and the inclination angle error of 6.42 degrees when the target is much smaller than the laser spot area (only 2.55 degrees o of the spot area). This study provides theoretical and technical support for the acquisition of target multi-dimensional information.

Reflective tomography lidar (RTL) bypasses the diffraction limit and atmospheric turbulence through non-coherent active illumination for long-range remote sensing, yet sparse angular sampling in dynamic space situational awareness triggers the "missing cone" problem, rendering inverse reconstruction ill-posed. In the paper, we propose a physics-informed dual-domain wavelet unrolling (DDWU) framework that couples wavelet-domain statistical regularization with projection-domain data consistency. Exploiting the sparsity gap between localized true edges and distributed aliasing streaks in high-frequency wavelet sub-bands, DDWU regularizes the ill-posed retrieval while preserving fine structural details. This hybrid architecture recovers high-fidelity structure from only 20% Nyquist sampling while suppressing turbulence and speckle artifacts. Field experiments at 10.38km validate 1.7cm resolution, establishing a robust methodology for non-cooperative long-range sensing that bridges data-driven restoration and physical measurement mechanisms.

Laser-induced breakdown spectroscopy (LIBS) technology has been widely applied across various fields due to its rapid and straightforward analytical capabilities. However, this technology is susceptible to noise interference during the detection process, which will seriously affect the quantitative analysis accuracy. To mitigate the influence of noise and improve the analysis accuracy, we propose a Gradient Histogram Constraint Truncated Weighted Nuclear Norm Minimization (GHCTWNNM) algorithm for LIBS spectra denoising. Here, we innovatively convert the denoising problem of 1D spectra data into a 2D image denoising problem, where we can take advantage of the superior image denoising technology to enhance the denoising effect of LIBS spectra. On the basis of the traditional WNNM algorithm, we introduce the truncation threshold and gradient histogram constraints, which not only improve the computational efficiency but also prevent distortion issues caused by excessive smoothing of image texture details. Subsequently, we derived the solution of the GHCTWNNM algorithm using the Alternating Direction Method of Multipliers (ADMM) method. The experimental results demonstrate that the GHCTWNNM algorithm achieves a remarkable improvement in denoising performance, with an increase of approximately 6 dB in Delta SNR compared to the WNNM algorithm. Moreover, in comparison with nine other image denoising algorithms, GHCTWNNM not only delivers superior denoising capabilities but also exhibits greater adaptability to different noise environments, especially in a high background noise environment. Additionally, the R2 of the Al element quantitative analysis result has increased by 0.26 after applying the GHCTWNNM denoising method. In summary, the LIBS denoising method based on the GHCTWNNM algorithm can effectively enhance the spectra SNR and significantly reduce the errors in quantitative analysis caused by noise, thereby enhancing the accuracy and reliability of LIBS. This provides a strong basis for its wide application and further development in various related fields.

The discovery of the vortex beam carrying orbital angular momentum (OAM) has made it a popular research object for enhancing the transmission capacity in free-space optical communication. However, the atmospheric turbulence effect in space has become a negative factor affecting the transmission quality of the vortex beam. Therefore, this paper proposes an anti-turbulence method of vortex beam OAM coding based on the trellis-coded modulation technique. Instead of traditional OAM coding of the bit stream, the source information is first modulated into constellation symbols after error correction coding, and then OAM coding is performed based on the constellation symbols and detected using a convolutional neural network. The experimental results confirm that the scheme proposed in this paper has a better bit error rate (BER) performance compared with the traditional scheme under different turbulence intensities, as well as more stable turbulence suppression at different transmission distances, and the BER performance shows a more significant improvement after increasing the number of samples, which provides a reference for studying the anti-turbulence effect of vortex beams based on coding techniques.

Despite the wide applications of full-waveform light detection and ranging (FW-LiDAR) on target detection and recognizing, topographical mapping, and ecological management, etc., the mapping between the echo waveform and the properties of the targets, even for typical three-dimensional (3D) targets, has not been established. The mechanics of the modulation of targets on the echo waveform is thus ambiguous, constraining the retrieval of target properties in FW-LiDAR. This paper derived the formula of echo waveform modulated by typical 3D targets, namely, a rectangular prism, a regular hexagonal prism, and a cone. The modulation of shape, size, position, and attitude of 3D targets on the echo waveform has been investigated extensively. The results showed that, for prisms, variations in the echo waveforms under various factors essentially arise from changes in the inclination angles of their reflective surfaces and their positions relative to the laser spot. For cones, their echo waveforms can be approximated and analyzed using isosceles triangular micro-facets. The work in this paper is helpful in probing the modulation of 3D targets on echo waveform, as well as extracting the properties of 3D targets in FW-LiDAR domains, which are significant in areas ranging from topographical mapping to space debris monitoring.

Accurate monitoring of trace wear metals in lubricants is critical for predictive maintenance of mechanical systems, yet remains challenging due to the inadequate sensitivity of conventional techniques with microliter samples. We demonstrate a novel strategy combining nanoparticle-enhanced laser-induced breakdown spectroscopy (NELIBS) with indirect ablation on metal substrates for ultrasensitive quantification of trace metals in engine oil. This approach overcomes fundamental limitations in viscous liquid analysis-including signal instability, matrix effects, and splashing-by leveraging laser ablation of conductive substrates coupled with localized surface plasmon resonance (LSPR) of gold nanoparticles. Crucially, metal substrates (Al) enable synergistic plasma enhancement where semiconductor substrates (Si) fail. Remarkably, the method achieves a 10fold signal enhancement for target wear metals (Mg), with detection limits of 0.36 ppm (Mg), 0.19 ppm (Ca), and 0.47 ppm (Ba), and excellent linearity (R2 > 0.99) validates quantification robustness. This non-contact technique provides a rapid, cost-effective solution for real-time engine health assessment, with significant implications for industrial oil monitoring and failure prediction.

We developed a novel quantum dots-enhanced laser-induced breakdown spectroscopy (QDE-LIBS) technique for the ultrasensitive detection of trace metals in lubricating oils. By utilizing AgNC@AgAux core-shell quantum dots (QDs) as signal amplifiers, this strategy overcomes matrix interference and reproducibility limitations inherent in typical LIBS. Predeposition of QDs on aluminum substrates generates a strong localized electric field and charge separation via quantum confinement effects, achieving a 21.8-fold signal enhancement at Mg II 279.553 nm. The signal-to-background ratio (SBR) at Ca II 393.367 nm is 10.65-fold higher than that at typical LIBS, and the relative standard deviation (RSD) at Ca II 393.367 nm is reduced by 71.6%. Quantitative analysis demonstrates excellent linearity (R2 > 0.97), with detection limits of 0.032 mg/L for Mg, 0.0656 mg/L for Ca, and 0.0882 mg/L for Ba. This method achieves exceptional sensitivity and reproducibility, with the potential for field deployment through automated sample preparation or precoated substrates, making it a promising candidate for wear metal monitoring in predictive maintenance systems.

Significance Hyperspectral lidar(HSL),an emerging active remote sensing technology,integrates the three-dimensional(3D)spatial detection capability of traditional lidar with the rich spectral information of hyperspectral imaging,addressing the long-standing limitation of separate spatial and spectral information acquisition in conventional remote sensing.Unlike passive hyperspectral imaging(which lacks 3D perception)and single-wavelength lidar(which lacks spectral discrimination),HSL simultaneously captures high-resolution 3D coordinates and spectral reflectance characteristics of targets,generating four-dimensional spatial-spectral point clouds.This unique capability is pivotal for advancing precision applications such as forest resource surveys(quantifying vertical structure and biochemical components),land cover classification(enhancing accuracy via spectral-spatial synergy),urban 3D modeling(distinguishing material properties),and target detection(penetrating obscurations).By enabling"one-stop"acquisition of both physical structure and chemical composition,HSL revolutionizes how we perceive and analyze complex environments,making it indispensable for addressing global challenges like sustainable resource management,and smart urban development. Progress Over the past two decades,HSL has evolved from dual-wavelength prototypes to sophisticated systems with tens even hundred spectral channels,driven by advancements in supercontinuum laser sources and spectral detection technologies.Key progress includes: 1)System architectures:Two dominant spectral splitting schemes have been developed:spatial splitting(using gratings for simultaneous multi-wavelength detection,suitable for airborne large-area scanning)and wavelength scanning[using acousto-optic tunable filter/liquid crystal tunable filter(AOTF/LCTF)for high spectral resolution,ideal for fine spectral analysis].Representative systems,such as the 56-channel airborne HSL(Wuhan University)have achieved detection ranges up to 500 m and 101-channel ground-based HSL(Anhui Jianzhu University),and spectral resolution as high as 5 nm. 2)Waveform processing:To extract accurate spatial-spectral information from overlapping echoes,methods like multi-spectral waveform decomposition(MSWD),multi-channel interconnection waveform decomposition(MIWD),and range resolution enhanced method with spectral properties(RREM)have been proposed.These techniques will enhance range resolution for lidar signals by leveraging cross-channel spectral correlations,overcoming the limitations of single-wavelength decomposition. 3)Radiometric correction:Strategies to mitigate distance effect(via piecewise fitting),incidence angle effect(using Lambertian-Beckmann models),and sub-footprint effect(through spectral ratio and area-weighted correction)have been developed,ensuring reliable spectral reflectance retrieval across diverse targets(vegetation,minerals,building materials). 4)Spatial-spectral point cloud applications:Techniques for point cloud generation(enabling true-color imaging without passive data),classification(combining machine/deep learning with spatial-spectral features),and feature extraction(e.g.,crop nitrogen content,mineral identification)have been validated,with classification accuracies exceeding 90%in vegetation and mineral scenarios. Conclusions and Prospects HSL has demonstrated significant potential in various applications,including vegetation monitoring,mineral exploration,and urban modeling,by providing detailed spatial and spectral information.However,challenges remain:limited detection range(mostly<100 m for ground systems),slow multi-channel data processing(lagging behind acquisition rates),and high system complexity hindering commercialization.Future research should focus on:1)System advancement:developing miniaturized,multi-platform(airborne,satellite-borne,underwater)systems via high-power supercontinuum lasers and low-loss spectral splitters to extend detection range and reduce cost;2)Information processing:enhancing real-time performance through hardware acceleration(FPGA/ASIC)and deep learning-based multi-channel waveform decomposition,and improving radiometric correction for non-Lambertian targets;3)Application expansion:exploring new frontiers such as defense reconnaissance(obscured target identification)and smart agriculture(3D biochemical mapping),supported by open datasets and standardized processing workflows.As these challenges are addressed,HSL is poised to become a cornerstone technology in high-precision remote sensing,enabling unprecedented insights into Earth systems and beyond.

High-accuracy quantitative analysis model of LIBS with small sample.

In this paper, we demonstrate an ultrafast diamond Raman laser at 1240 nm with a pulse repetition rate of 956.62 MHz, a pulse duration of 39.5 ps, and an average power of up to 3.3 W based on synchronous pumping. The pump source is an electrical-pulse-modulated picosecond pulsed laser at 1064 nm with a repetition rate of 239.16 MHz and a pulse duration of 65.4 ps. A quadrupling repetition rate of the Raman pulse is achieved by synchronously amplifying both the forward and backward Raman pulses and the amplified Raman pulse undergoing two round trips in the resonator within one pump pulse period. The compression ratio of the pulse duration from the pump to the Raman is 1.66. This work offers a convenient and efficient method to significantly enhance the repetition rate of ultrafast crystalline Raman lasers and proves that a non-coherent ultrafast pump pulsed laser can be converted to a coherent mode-locked ultrafast Raman pulsed laser based on Raman conversion.

Photon counting lidar has emerged as a strong candidate technology for active detection applications because of its advantages of single photon sensitivity and high ranging accuracy. The timing histogram of a single pixel for photon counting lidar contains the target’s range information, while the laser echo of full-waveform lidar contains abundant structure and reflection information of the target. Based on the previous work of full waveform correction for stationary target, we propose a new method of the full waveform recovery for moving target, aiming at the issue of obtaining the characteristics of ultra-long-range moving targets under high-flux conditions. Our method achieves full-waveform recovery by means of data preprocessing, motion compensation, and photon waveform correction. Through simulation calculations, we analyze and compare the effectiveness of each step of the method. Compared with the raw histogram, the Normalized Root Mean Square Error (NRMSE) of the recovery full waveform and the ideal waveform is reduced from 0.137 to 0.032. Furthermore, we validate the algorithm’s robustness. As the speed increases from 5 to 340m/s, the NRMSE is always less than 0.04. The results indicate that the recovery waveform of targets hardly vary with changes in velocity. For an accumulation of 200 pulses, when the signal photons is 0.019 − 3 and the signal-to-noise ratio is below 0.033, the algorithm consistently exhibits excellent performance. Besides, we have demonstrated that for single-layer moving targets, multi-layer moving targets, and round-trip moving targets, the algorithm has good performance on the recovery of the targets’ full-waveform, and the NRMSE is less than 0.0054. This provides a new idea for obtaining the shape of targets with variable speeds at a single pixel, and provides exciting news for applications such as detection and recognition of ultra-long-range aerial targets and detection of space debris.

Dual-comb ranging (DCR), with its superior overall performance compared to traditional ranging technologies, has recently attracted widespread interest in the research community. Nevertheless, the ranging distance or the material of the targets is limited by the detection sensitivity of optical asynchronous linear sampling. This limitation restricts the application of DCR in several highly significant scenarios. Here, we utilize the photon-counting method to dramatically break through the detection sensitivity to femtowatt. To overcome the impact of fiber-length wandering and achieve Michelson interference based absolute distance measurement, an orthogonal polarization interferometry-arm configuration and a reference-arm based photon-counting trigger protocol are proposed. This photon-counting DCR system can conduct long-period photon-counting coherently, thus, realizing the lowest detection power of phase-stabled DCR to date. The results show that with only 18 femtowatt average power detected, the time-of-flight and multi-wavelength interferometry yields a precision of 22 µm and 8 nm in 3 min, respectively. This work paves the way for the field of large-scale spacecraft formation flying, synthetic aperture space telescope position attitude control, interplanetary positioning, and hard target distance measurement.

Single-photon lidar stands out as a promising technology for long-distance lidar applications, owing to its attributes of single-photon sensitivity and high repetition rate. Existing single-photon lidar systems typically rely on single-point scanning for positioning and tracking, necessitating intricate and precise scanning control. In pursuit of a more concise and efficient positioning, we incorporate the four-quadrant theory to articulate the signal formula of photon detection, and propose a novel single-photon four-quadrant positioning method. Our method, which includes signal preprocessing, compensation for longitudinal motion, extraction of pixel intensity, and acquisition of lateral motion, facilitates motion acquisition and positioning for targets. Through simulation calculations, we analyze and compare the effectiveness of each step of the method. With longitudinal and lateral speeds of 100 m/s and 50 m/s, respectively, the trajectory error is 1.7%, and the average speed error is 1.8%. Moreover, for various verification experiments, the trajectory errors are all below 4.2%, and the average speed errors remain under 5.4%, effectively verifying the validity of our method in acquiring the motion information and positioning of targets. It provides an excellent option for acquiring motion information and tracking small moving targets over long distances.

Photon counting lidar has revolutionized the field of lidar technology with its exceptional single-photon sensitivity and picosecond-level time resolution. It is particularly effective for detecting ultra-long-range targets and measuring global ecosystems. Timing histograms and waveforms play vital roles in these applications, as they contain rich structural information about the targets. To systematically explore the performance boundary model for full-waveform applications in photon counting lidar, we have developed a model based on underlying theory and feasibility, which overcomes the limitation of detecting fewer than 5% of illumination cycles. By using the cumulative emission pulse number as the objective function, we establish the performance boundary model for full-waveform in photon counting lidar, revealing the relationship between the accuracy of the full-waveform and system parameters. The model’s accuracy is verified through theoretical analysis and experimental validation. Subsequently, we utilize Pareto Optimality to determine the optimal parameters for the full-waveform performance boundary model. Experimental data indicates that, to ensure a normalized root mean square error ( nRMSE ) of less than 0.03 between full-waveform and ideal waveform, the performance boundaries are as follows: the optimal time bin width is 256ps, the signal intensity falls within [0.8, 1.6], the tolerable noise is [0, 0.63M]Hz, and the minimum cumulative pulses required is between [282, 319], given that the echo width is 5ns. Finally, we discuss the practical application of the full-waveform performance boundary in photon counting lidar for complex target detection scenarios. Under the optimal parameter configuration, the R-Square ( R 2 ) between the full-waveform and the ideal waveform consistently exceeds 90%. This work not only expands the range of applications for photon counting lidar in the field of full-waveform, but also establishes a strong connection with full-waveform processing algorithms.