Mangrove monitoring and extraction based on multi-source remote sensing data:a deep learning method based on SAR and optical image fusion

Mangroves are indispensable to coastlines,maintaining biodiversity,and mitigating climate change.Therefore,improving the accuracy of mangrove information identification is crucial for their ecological protection.Aiming at the limited morphological information of synthetic aperture radar(SAR)images,which is greatly interfered by noise,and the susceptibility of optical images to weather and lighting conditions,this paper proposes a pixel-level weighted fusion method for SAR and optical images.Image fusion enhanced the target features and made mangrove monitoring more comprehensive and accurate.To address the problem of high similarity between mangrove forests and other forests,this paper is based on the U-Net convolutional neural network,and an attention mechanism is added in the feature extraction stage to make the model pay more attention to the mangrove vegetation area in the image.In order to accelerate the convergence and normalize the input,batch normalization(BN)layer and Dropout layer are added after each convolutional layer.Since mangroves are a minority class in the image,an improved cross-entropy loss function is introduced in this paper to improve the model’s ability to recognize mangroves.The AttU-Net model for mangrove recognition in high similarity environments is thus constructed based on the fused images.Through comparison experiments,the overall accuracy of the improved U-Net model trained from the fused images to recognize the predicted regions is significantly improved.Based on the fused images,the recognition results of the AttU-Net model proposed in this paper are compared with its benchmark model,U-Net,and the Dense-Net,Res-Net,and Seg-Net methods.The AttU-Net model captured mangroves’complex structures and textural features in images more effectively.The average OA,F1-score,and Kappa coefficient in the four tested regions were 94.406%,90.006%,and 84.045%,which were significantly higher than several other methods.This method can provide some technical support for the monitoring and protection of mangrove ecosystems.

Experimental Study on the Variation of Optical Imaging Characteristics with Zenith Angle due to Internal Solitary Waves in Sunglint

Internal solitary waves(ISWs)change the roughness of the sea surface,thus producing dark and bright bands in optical images.However,reasons for changes in imaging characteristics with the solar zenith angle remain unclear.In this paper,the optical imaging pattern of ISWs in sunglint under different zenith angles of the light source is investigated by collecting optical images of ISWs through physical simulation.The experiment involves setting 10 zenith angles of the light source,which are divided into area a the optical images of ISWs in the three areas show dark-bright mode,single bright band,and bright-dark mode,which are consistent with those observed by optical remote sensing.In addition,this study analyzed the percentage of the dark and bright areas of the bands and the change in the relative gray difference and found changes in both areas under different zenith angles of the light source.The MODIS and ASAR images display a similar brightness-darkness distance of the same ISWs.Therefore,the relationship between the brightness-darkness distance and the characteristic half-width of ISWs is determined in accordance with the eKdV theory and the imaging mechanism of ISWs of the SAR image.Overall,the relationship between them in the experiment is almost consistent with the theoretical result.

Reconstruction algorithm for cross-waveband optical computing imaging

In a single-pixel fast imaging setup,the data collected by the single-pixel detector needs to be processed by a computer,but the speed of the latter will affect the image reconstruction time.Here we propose two kinds of setups which are able to transform non-visible into visible light imaging,wherein their computing process is replaced by a camera integration mode.The image captured by the camera has a low contrast,so here we present an algorithm that can realize a high quality image in near-infrared to visible cross-waveband imaging.The scheme is verified both by simulation and in actual experiments.The setups demonstrate the great potential for single-pixel imaging and high-speed cross-waveband imaging for future practical applications.

Detecting early changes in choroidal vascularity and thickness using optical coherence tomography in patients with corneal crosslinking for keratoconus

AIM:To investigate changes in choroidal thickness and vascularity in keratoconus patients treated with corneal crosslinking.METHODS:This study evaluated 28 eyes of 22 patients with keratoconus who underwent corneal crosslinking.The choroidal thicknesses were evaluated on enhanced depth imaging optical coherence tomography at the preoperative and postoperative 3d,1,and 3mo.Choroidal thickness in the four cardinal quadrants and the fovea were evaluated.The choroidal vascularity index was also calculated.RESULTS:There was no significant difference in central choroidal thickness between the preoperative and postoperative 3d,1mo(P>0.05).There was a significant increase in the 3mo(P=0.034)and a significant decrease in the horizontal choroidal vascularity index on the postoperative 3d(P=0.014),there was no statistically significant change in vertical axes and other visits in horizontal sections(P>0.05).CONCLUSION:This study sheds light on choroidal changes in postoperative corneal crosslinking for keratoconus.While it suggests the procedure’s relative safety for submacular choroid,more extensive research is necessary to confirm these findings and their clinical significance.

The Application of Digital Optical 3D Image Analyzer Evaskin in The Evaluation of Wrinkles

To verify the effectiveness of digital optical 3D image analyzer EvaSKIN in the objective and quantitative evaluation of wrinkles.A total of 115 subjects were recruited,the facial images of the subjects were collected by digital optical 3D image analyzer and manual camera,the changes of crow’s feet with age were analyzed.Pictures obtained by manual photography can be directly used for observation and preliminary grading of wrinkles.However,the requirements for evaluators are high,and the results are prone to errors,which will affect the accuracy of the evaluation.Therefore,skilled raters are needed.Compared with the manual photography method,the digital optical 3D image analyzer EvaSKIN can realize three-dimensional extraction of wrinkles,and obtain the change trend of crow’s feet with age.20~30 years old,wrinkles begin to appear slowly;wrinkles will increase rapidly at the age of 30~50;The length of 50~60 year old wrinkles is basically fixed,the wrinkles develop longitudewise,gradually widen and deepen,and the area,depth and volume increase is obvious,and the skin aging condition is intensified.the digital optical 3D image analyzer EvaSKIN realizes the 3D extraction of wrinkles,quantifies the circumference,area,average depth,maximum depth and volume of wrinkles,realizes the objective and quantitative evaluation of wrinkle state,is more accurate in the measurement of wrinkles,and provides a new instrument and method for the evaluation of wrinkles.it is a perfect and supplement to the traditional evaluation methods,and to a certain extent,it helps the research and development and evaluation institutions of cosmetics to obtain more abundant and three-dimensional data support.

Mapping the bathymetry of shallow coastal water using singleframe fine-resolution optical remote sensing imagery

This paper presents a bathymetry inversion method using single-frame fine-resolution optical remote sensing imagery based on ocean-wave refraction and shallow-water wave theory. First, the relationship among water depth, wavelength and wave radian frequency in shallow water was deduced based on shallow-water wave theory. Considering the complex wave distribution in the optical remote sensing imagery, Fast Fourier Transform （FFT） and spatial profile measurements were applied for measuring the wavelengths. Then, the wave radian frequency was calculated by analyzing the long-distance fluctuation in the wavelength, which solved a key problem in obtaining the wave radian frequency in a single-frame image. A case study was conducted for Sanya Bay of Hainan Island, China. Single-flame fine-resolution optical remote sensing imagery from QuickBird satellite was used to invert the bathymetry without external input parameters. The result of the digital elevation model （DEM） was evaluated against a sea chart with a scale of 1：25 000. The root-mean-square error of the inverted bathymetry was 1.07 m, and the relative error was 16.2%. Therefore, the proposed method has the advantages including no requirement for true depths and environmental parameters, and is feasible for mapping the bathymetry of shallow coastal water.

Cerebrovascular segmentation from mesoscopic optical images using Swin Transformer

Vascular segmentation is a crucial task in biomedical image processing,which is significant for analyzing and modeling vascular networks under physiological and pathological states.With advances in fluorescent labeling and mesoscopic optical techniques,it has become possible to map the whole-mouse-brain vascular networks at capillary resolution.However,segmenting vessels from mesoscopic optical images is a challenging task.The problems,such as vascular signal discontinuities,vessel lumens,and background fluorescence signals in mesoscopic optical images,belong to global semantic information during vascular segmentation.Traditional vascular segmentation methods based on convolutional neural networks(CNNs)have been limited by their insufficient receptive fields,making it challenging to capture global semantic information of vessels and resulting in inaccurate segmentation results.Here,we propose SegVesseler,a vascular segmentation method based on Swin Transformer.SegVesseler adopts 3D Swin Transformer blocks to extract global contextual information in 3D images.This approach is able to maintain the connectivity and topology of blood vessels during segmentation.We evaluated the performance of our method on mouse cerebrovascular datasets generated from three different labeling and imaging modalities.The experimental results demonstrate that the segmentation effect of our method is significantly better than traditional CNNs and achieves state-of-the-art performance.

Wide-spectrum optical synthetic aperture imaging via spatial intensity interferometry

High resolution imaging is achieved using increasingly larger apertures and successively shorter wavelengths.Optical aperture synthesis is an important high-resolution imaging technology used in astronomy.Conventional long baseline amplitude interferometry is susceptible to uncontrollable phase fluctuations,and the technical difficulty increases rapidly as the wavelength decreases.The intensity interferometry inspired by HBT experiment is essentially insensitive to phase fluctuations,but suffers from a narrow spectral bandwidth which results in a lack of effective photons.In this study,we propose optical synthetic aperture imaging based on spatial intensity interferometry.This not only realizes diffraction-limited optical aperture synthesis in a single shot,but also enables imaging with a wide spectral bandwidth,which greatly improves the optical energy efficiency of intensity interferometry.And this method is insensitive to the optical path difference between the sub-apertures.Simulations and experiments present optical aperture synthesis diffraction-limited imaging through spatial intensity interferometry in a 100 nm spectral width of visible light,whose maximum optical path difference between the sub-apertures reaches 69λ.This technique is expected to provide a solution for optical aperture synthesis over kilometer-long baselines at optical wavelengths.

Measurement of optical coherence structures of random optical fields using generalized Arago spot experiment

The optical coherence structures of random optical fields can determine beam propagation behavior,light–matter interactions,etc.Their performance makes a light beam robust against turbulence,scattering,and distortion.Recently,we proposed optical coherence encryption and robust far-field optical imaging techniques.All related applications place a high demand on precision in the experimental measurements of complex optical coherence structures,including their real and imaginary parts.Past studies on these measurements have mainly adopted theoretical mathematical approximations,limited to Gaussian statistic involving speckle statistic(time-consuming),or used complicated and delicate optical systems in the laboratory.In this study,we provide:a robust,convenient,and fast protocol to measure the optical coherence structures of random optical fields via generalized Arago(or Poisson)spot experiments with rigorous mathematical solutions.Our proposal only requires to capture the intensity thrice,and is applicable to any optical coherence structures,regardless of their type or optical statistics.The theoretical and experimental results demonstrated that the real and imaginary parts of the structures could be simultaneously recovered with high precision.We believe that such a protocol can be widely employed in phase measurement,optical imaging,and image transfer.

Optical image encryption algorithm based on a new four-dimensional memristive hyperchaotic system and compressed sensing

Some existing image encryption schemes use simple low-dimensional chaotic systems, which makes the algorithms insecure and vulnerable to brute force attacks and cracking. Some algorithms have issues such as weak correlation with plaintext images, poor image reconstruction quality, and low efficiency in transmission and storage. To solve these issues,this paper proposes an optical image encryption algorithm based on a new four-dimensional memristive hyperchaotic system(4D MHS) and compressed sensing(CS). Firstly, this paper proposes a new 4D MHS, which has larger key space, richer dynamic behavior, and more complex hyperchaotic characteristics. The introduction of CS can reduce the image size and the transmission burden of hardware devices. The introduction of double random phase encoding(DRPE) enables this algorithm has the ability of parallel data processing and multi-dimensional coding space, and the hyperchaotic characteristics of 4D MHS make up for the nonlinear deficiency of DRPE. Secondly, a construction method of the deterministic chaotic measurement matrix(DCMM) is proposed. Using DCMM can not only save a lot of transmission bandwidth and storage space, but also ensure good quality of reconstructed images. Thirdly, the confusion method and diffusion method proposed are related to plaintext images, which require both four hyperchaotic sequences of 4D MHS and row and column keys based on plaintext images. The generation process of hyperchaotic sequences is closely related to the hash value of plaintext images. Therefore, this algorithm has high sensitivity to plaintext images. The experimental testing and comparative analysis results show that proposed algorithm has good security and effectiveness.

Real-time and high-transmission middle-infrared optical imaging system based on a pixel-wise metasurface micro-polarization array

Real-time polarization medium-wave infrared(MIR)optical imaging systems enable the acquisition of infrared and polarization information for a target.At present,real-time polarization MIR devices face the following problems:poor real-time performance,low transmission and high requirements for fabrication and integration.Herein,we aim to improve the performance of real-time polarization imaging systems in the MIR waveband and solve the above-mentioned defects.Therefore,we propose a MIR polarization imaging system to achieve real-time polarization-modulated imaging with high transmission as well as improved performance based on a pixel-wise metasurface micro-polarization array(PMMPA).The PMMPA element comprises several linear polarization(LP)filters with different polarization angles.The optimization results demonstrate that the transmittance of the center field of view for the LP filters is up to 77%at a wavelength of4.0μm and an extinction ratio of 88 d B.In addition,a near-diffraction-limited real-time MIR imaging optical system is designed with a field of view of 5°and an F-number of 2.The simulation results show that an MIR polarization imaging system with excellent real-time performance and high transmission is achieved by using the optimized PMMPA element.Therefore,the method is compatible with the available optical system design technologies and provides a way to realize real-time polarization imaging in MIR wavebands.

Observing single cells in whole organs with optical imaging

Cells are the basic unit of human organs that are not fully understood.The revolutionary advancements of optical imaging alowed us to observe single cells in whole organs,revealing the complicated composition of cells with spatial information.Therefore,in this review,we revisit the principles of optical contrast related to those biomolecules and the optical techniques that transform optical contrast into detectable optical signals.Then,we describe optical imaging to achieve threedimensional spatial discrimination for biological tisutes.Due to the milky appearance of tissues,the spatial information burred deep in the whole organ.Fortunately,strategies developed in the last decade could circumvent this issue and lead us into a new era of investigation of the cells with their original spatial information.

Research on the model of high robustness computational optical imaging system

Computational optical imaging is an interdisciplinary subject integrating optics, mathematics, and information technology. It introduces information processing into optical imaging and combines it with intelligent computing, subverting the imaging mechanism of traditional optical imaging which only relies on orderly information transmission. To meet the high-precision requirements of traditional optical imaging for optical processing and adjustment, as well as to solve its problems of being sensitive to gravity and temperature in use, we establish an optical imaging system model from the perspective of computational optical imaging and studies how to design and solve the imaging consistency problem of optical system under the influence of gravity, thermal effect, stress, and other external environment to build a high robustness optical system. The results show that the high robustness interval of the optical system exists and can effectively reduce the sensitivity of the optical system to the disturbance of each link, thus realizing the high robustness of optical imaging.

An Intelligent Detection Method for Optical Remote Sensing Images Based on Improved YOLOv7

To address the issue of imbalanced detection performance and detection speed in current mainstream object detection algorithms for optical remote sensing images,this paper proposes a multi-scale object detection model for remote sensing images on complex backgrounds,called DI-YOLO,based on You Only Look Once v7-tiny(YOLOv7-tiny).Firstly,to enhance the model’s ability to capture irregular-shaped objects and deformation features,as well as to extract high-level semantic information,deformable convolutions are used to replace standard convolutions in the original model.Secondly,a Content Coordination Attention Feature Pyramid Network(CCA-FPN)structure is designed to replace the Neck part of the original model,which can further perceive relationships between different pixels,reduce feature loss in remote sensing images,and improve the overall model’s ability to detect multi-scale objects.Thirdly,an Implicitly Efficient Decoupled Head(IEDH)is proposed to increase the model’s flexibility,making it more adaptable to complex detection tasks in various scenarios.Finally,the Smoothed Intersection over Union(SIoU)loss function replaces the Complete Intersection over Union(CIoU)loss function in the original model,resulting in more accurate prediction of bounding boxes and continuous model optimization.Experimental results on the High-Resolution Remote Sensing Detection(HRRSD)dataset demonstrate that the proposed DI-YOLO model outperforms mainstream target detection algorithms in terms of mean Average Precision(mAP)for optical remote sensing image detection.Furthermore,it achieves Frames Per Second(FPS)of 138.9,meeting fast and accurate detection requirements.

Optical coherence tomography enhanced depth imaging of chorioretinal folds in patients with orbital tumors

AIM:To characterize spectral-domain optical coherence tomography(SD-OCT)features of chorioretinal folds in orbital mass imaged using enhanced depth imaging(EDI).METHODS:Prospective observational case-control study was conducted in 20 eyes of 20 patients,the uninvolved eye served as a control.All the patients underwent clinical fundus photography,computed tomography,EDI SDOCT imaging before and after surgery.Two patients with cavernous hemangiomas underwent intratumoral injection of bleomycin A5;the remaining patients underwent tumor excision.Patients were followed 1 to 14mo following surgery(average follow up,5.8mo).RESULTS:Visual acuity prior to surgery ranged from 20/20 to 20/200.Following surgery,5 patients’visual acuity remained unchanged while the remaining 15 patients had a mean letter improvement of 10(range 4 to 26 letters).Photoreceptor inner/outer segment defects were found in 10 of 15 patients prior to surgery.Following surgical excision,photoreceptor inner/outer segment defects fully resolved in 8 of these 10 patients.CONCLUSION:Persistence of photoreceptor inner/outer segment defects caused by compression of the globe by an orbital mass can be associated with reduced visual prognosis.Our findings suggest that photoreceptor inner/outer segment defects on EDI SD-OCT could be an indicator for immediate surgical excision of an orbital mass causing choroidal compression.

Imaging assessment of photosensitizer emission induced by radionuclide-derived Cherenkov radiation using charge-coupled device optical imaging and long-pass filters

BACKGROUND Radionuclides produce Cherenkov radiation(CR),which can potentially activate photosensitizers(PSs)in phototherapy.Several groups have studied Cherenkov energy transfer to PSs using optical imaging;however,cost-effectively identifying whether PSs are excited by radionuclide-derived CR and detecting fluorescence emission from excited PSs remain a challenge.Many laboratories face the need for expensive dedicated equipment.AIM To cost-effectively confirm whether PSs are excited by radionuclide-derived CR and distinguish fluorescence emission from excited PSs.METHODS The absorbance and fluorescence spectra of PSs were measured using a microplate reader and fluorescence spectrometer to examine the photo-physical properties of PSs.To mitigate the need for expensive dedicated equipment and achieve the aim of the study,we developed a method that utilizes a chargecoupled device optical imaging system and appropriate long-pass filters of different wavelengths(manual sequential application of long-pass filters of 515,580,645,700,750,and 800 nm).Tetrakis(4-carboxyphenyl)porphyrin(TCPP)was utilized as a model PS.Different doses of copper-64(^(64)CuCl_(2))(4,2,and 1 mCi)were used as CR-producing radionuclides.Imaging and data acquisition were performed 0.5 h after sample preparation.Differential image analysis was conducted by using ImageJ software(National Institutes of Health)to visually evaluate TCPP fluorescence.RESULTS The maximum absorbance of TCPP was at 390-430 nm,and the emission peak was at 670 nm.The CR and CRinduced TCPP emissions were observed using the optical imaging system and the high-transmittance long-pass filters described above.The emission spectra of TCPP with a peak in the 645-700 nm window were obtained by calculation and subtraction based on the serial signal intensity(total flux)difference between^(64)CuCl_(2)+TCPP and^(64)CuCl_(2).Moreover,the differential fluorescence images of TCPP were obtained by subtracting the^(64)CuCl_(2)image from the^(64)CuCl_(2)+TCPP image.The experimental results considering different^(64)CuCl_(2)doses showed a dosedependent trend.These results demonstrate that a bioluminescence imaging device coupled with different longpass filters and subtraction image processing can confirm the emission spectra and differential fluorescence images of CR-induced TCPP.CONCLUSION This simple method identifies the PS fluorescence emission generated by radionuclide-derived CR and can contribute to accelerating the development of Cherenkov energy transfer imaging and the discovery of new PSs.

NIR-II fluorescence imaging in liver tumor surgery: A narrative review

In liver tumor surgery,the recognition of tumor margin and radical resection of microcancer focis have always been the crucial points to reduce postoperative recurrence of tumor.However,naked-eye inspection and palpation have limited effectiveness in identifying tumor boundaries,and traditional imaging techniques cannot consistently locate tumors in real time.As an intraoperative real-time navigation imaging method,NIRfluorescence imaging has been extensively studied for its simplicity,reliable safety,and superior sensitivity,and is expected to improve the accuracy of liver tumor surgery.In recent years,the research focus of NIRfluorescence has gradually shifted from the-rst near-infrared window(NIR-I,700–900 nm)to the second near-infrared window(NIR-II,1000–1700 nm).Fluorescence imaging in NIR-II reduces the scattering effect of deep tissue,providing a preferable detection depth and spatial resolution while signi-cantly eliminating liver autofluorescence background to clarify tumor margin.Developingfluorophores combined with tumor antibodies will further improve the precision offluorescence-guided surgical navigation.With the development of a bunch offluorophores with phototherapy ability,NIR-II can integrate tumor detection and treatment to explore a new therapeutic strategy for liver cancer.Here,we review the recent progress of NIR-IIfluorescence technology in liver tumor surgery and discuss its challenges and potential development direction.

Opto-mechanical-thermal integration analysis of Doppler asymmetric spatial heterodyne interferometer

In order to improve the detection accuracy of Doppler asymmetric spatial heterodyne(DASH)interferometer in harsh temperatures,an opto-mechanical-thermal integration analysis is carried out.Firstly,the correlation between the interference phase and temperature is established according to the working principle and the phase algorithm of the interferometer.Secondly,the optical mechanical thermal analysis model and thermal deformation data acquisition model are designed.The deformation data of the interference module and the imaging optical system at different temperatures are given by temperature load simulation analysis,and the phase error caused by thermal deformation is obtained by fitting.Finally,based on the wind speed error caused by thermal deformation of each component,a reasonable temperature control scheme is proposed.The results show that the interference module occupies the main cause,the temperature must be controlled within(20±0.05)℃,and the temperature control should be carried out for the temperature sensitive parts,and the wind speed error caused by the part is 3.8 m/s.The thermal drift between the magnification of the imaging optical system and the thermal drift of the relative position between the imaging optical system and the detector should occupy the secondary cause,which should be controlled within(20±2)℃,and the wind speed error caused by the part is 3.05 m/s.In summary,the wind measurement error caused by interference module,imaging optical system,and the relative position between the imaging optical system and the detector can be controlled within 6.85 m/s.The analysis and temperature control schemes presented in this paper can provide theoretical basis for DASH interferometer engineering applications.

Clinical characteristics of peripapillary hyperreflective ovoid mass-like structures in myopic children

AIM:To describe the characteristics of peripapillary hyperreflective ovoid mass-like structure(PHOMS)in myopic children and to investigate factors associated with PHOMS.METHODS:This retrospective observational study included 101 eyes of 101 children(age≤17y)with myopia.All included patients underwent comprehensive clinical examination.Optic nerve canal parameters,including disc diameter,optic nerve head(ONH)tilt angle,and border tissue angle were measured using serial enhanced-depth imaging spectral-domain optical coherence tomography(EDI-OCT).Based on the optic disc drusen consortium’s definition of PHOMS,eyes were classified as PHOMS group and non-PHOMS group.PHOMS was categorized according to height.RESULTS:Sixty-seven(66.3%)eyes were found with PHOMS.Small PHOMS could only be detected by optical coherence tomography(OCT).Medium PHOMS could be seen with blurred optic disc borders corresponding to OCT.The most frequent location of PHOMS was at the nasosuperior(91%,61 of 67 eyes)to ONH disc.The axial length and spherical equivalent were more myopic in the PHOMS group than in the non-PHOMS group(both P<0.001).ONH tilt angle was also significantly greater in PHOMS group than in non-PHOMS group[8.90(7.16-10.54)vs 3.93(3.09-5.25),P<0.001].Border tissue angle was significantly smaller in PHOMS group than in non-PHOMS group[29.70(20.90-43.81)vs 45.62(35.18-60.45),P<0.001].In the multivariable analysis,spherical equivalent(OR=3.246,95%CI=1.209-8.718,P=0.019)and ONH tilt angle(OR=3.275,95%CI=1.422-7.542,P=0.005)were significantly correlated with PHOMS.There was no disc diameter associated with PHOMS.In the linear regression analysis,border tissue angle was negatively associated with PHOMS height(β=-2.227,P<0.001).CONCLUSION:PHOMS is associated with optic disc tilt and optic disc nasal shift in myopia.Disc diameter is not a risk factor for PHOMS.The changes in ONH caused by axial elongation facilitated an understanding of the mechanism of PHOMS.

Tree species classification using deep learning and RGB optical images obtained by an unmanned aerial vehicle

The diversity of tree species and the complexity of land use in cities create challenging issues for tree species classification.The combination of deep learning methods and RGB optical images obtained by unmanned aerial vehicles(UAVs) provides a new research direction for urban tree species classification.We proposed an RGB optical image dataset with 10 urban tree species,termed TCC10,which is a benchmark for tree canopy classification(TCC).TCC10 dataset contains two types of data:tree canopy images with simple backgrounds and those with complex backgrounds.The objective was to examine the possibility of using deep learning methods(AlexNet,VGG-16,and ResNet-50) for individual tree species classification.The results of convolutional neural networks(CNNs) were compared with those of K-nearest neighbor(KNN) and BP neural network.Our results demonstrated:(1) ResNet-50 achieved an overall accuracy(OA) of 92.6% and a kappa coefficient of 0.91 for tree species classification on TCC10 and outperformed AlexNet and VGG-16.(2) The classification accuracy of KNN and BP neural network was less than70%,while the accuracy of CNNs was relatively higher.(3)The classification accuracy of tree canopy images with complex backgrounds was lower than that for images with simple backgrounds.For the deciduous tree species in TCC10,the classification accuracy of ResNet-50 was higher in summer than that in autumn.Therefore,the deep learning is effective for urban tree species classification using RGB optical images.