Imaging Mapping Research Articles

CLFNet: a multi-modal data fusion network for traffic sign extraction

Abstract When using image data for signage extraction, poor visibility conditions such as insufficient light, rainy days, and low light intensity are encountered, leading to low accuracy and poor boundary segmentation in vision-based detection methods. To address this problem, we propose a cross-modal latent feature fusion network for signage detection, which obtains rich boundary information by combining images with light detection and ranging depth images, thus compensating for the pseudo-boundary phenomenon that may occur when using a single RGB image segmentation. First, HRNet is utilized as the backbone network to extract the boundary information of the point cloud depth map and RGB image by introducing the boundary extraction module; Second, the sensitivity to the boundary is enhanced by applying the feature aggregation module to deeply fuse the extracted boundary information with the image features; Finally, boundary Intersection over Union (IOU) is introduced as an evaluation index. The results show that the method performs more superiorly compared to the mainstream RGBD network, with an improvement of 5.5% and 6.1% in IOU and boundary IOU, and an accuracy of 98.3% and 96.2%, respectively, relative to the baseline network.

XTC-DDPM: Diffusion Model for Tone Mapping and Pseudo-Color Imaging of Dual-Energy X-Ray Security Inspection Images

XTC-DDPM: Diffusion Model for Tone Mapping and Pseudo-Color Imaging of Dual-Energy X-Ray Security Inspection Images

MnOOH nanorods decorated with CeO2 nanoparticles as advanced electrode for high-performance supercapacitor

The development of novel composite electrode materials is essential to fabricating supercapacitors with high specific capacitance and good stability. In this study, MnOOH nanorods adorned with CeO2 (CeMn composites) have been satisfactorily synthesized through in-situ growth of tiny CeO2 nanoparticles using hydrothermal treatment. SEM images revealed that the granular CeO2 particles are adhered to the surfaces of nanorod-shaped MnOOH. XRD analysis confirmed the CeMn composites maintain the crystal structure of MnOOH and CeO2 with high purity. The EDS elemental mapping images demonstrated that Mn, O, and Ce elements are homogenously dispersion distributed in the CeMn composites. The supercapacitive performance of the MnOOH and CeMn composites pasted onto the Ni foam was evaluated determined through electrochemical measurements. The Ce0.05Mn1 (Ce/Mn molar ratio of 0.05/1) as a supercapacitor electrode exhibited an excellent specific capacitance of 857.62 F/g at 1 A/g, which is higher than the values for the MnOOH. Moreover, the prepared Ce0.05Mn1 still could retain good cycling stability over 3000 charge/discharge cycles. This study presents a feasible route to develop high-performing supercapacitor electrode materials.

Design of MnO2/g-C3N4 heterojunction composite photocatalysts for augmented charge separation and photocatalytic degradation performance with superior antibacterial activity

The development of advanced photocatalysts is critical for addressing environmental pollution and enhancing water purification processes. Our study has effectively developed a novel MnO2/g-C3N4 (MGC) heterojunction composite photocatalyst exhibiting superior photo-degradation under visible-light (VL) conditions and also established antibacterial activities. Comprehensive characterization was acquired using XRD, FT-IR, FE-SEM with EDX-associated mapping images, TEM, UV–Vis DRS and PL investigations, indicating the successful formation of well-dispersed MnO2 nanoparticles (NPs) over the g-C3N4 catalyst. The MGC composite heterojunction photocatalyst demonstrated increased photocatalytic degradation activity of 77.2 % in aqueous crystal violet (CV) under VL within 120 min, significantly outperforming pure g-C3N4 and MnO2 by 3.02 and 2.37 times, respectively. The MGC composite demonstrates remarkable stability and reusability, retaining 73.7 % of its efficiency after five consecutive cycles. Additionally, the as-synthesised composite endows potent antibacterial action against various pathogenic bacteria including K. pneumonia, S. aureus, E. coli and B. cereus. The active species analysis indicates that the composite photocatalyst facilitates charge transfer, while effectively preventing the recombination of photo-produced carriers via an effective Z-scheme mechanism, and the synergistic things among MnO2 and g-C3N4 are accredited to the boosted photocatalytic and antibacterial activities. This research describes a photocatalytic approach for efficiently eliminating various contaminants from water bodies, which has significant implications for the future development of photocatalytic technology for wastewater handling.

First Hit the Best Hit for Advanced Differentiated Papillary Thyroid Carcinoma?

Optimal upfront surgery with precise and objective image mapping is perhaps the way to go to enhance treatment outcomes and mitigate the use of post-operative radioiodine therapy. With the advent of targeted agents and precision medicine, it is now possible to undertake surgery for very advanced or inoperable differentiated thyroid cancers, followed by appropriate adjuvant treatment.

DeepUrbanMapper - Satellite Image to Google Maps Translation

Abstract: The research presents DeepUrbanMapper, an innovative framework developed for the translation of satellite imagery into Google Maps-style renderings utilizing advanced image-to-image translation techniques. It is based on Generative Adversarial Networks (GANs), widely known in machine learning for their ability to yield high-quality images. To train it, the process involves a carefully curated and labeled dataset that consists of paired satellite and map images, enabling DeepUrbanMapper to learn the complex mappings between these two distinct visual domains. The GAN architecture used in DeepUrbanMapper is carefully designed and optimized to preserve spatial coherence and improve visual quality of the translated image. This helps in producing output maps that are close to the ground truth in both visual and geographical aspects. The proposed method includes few novel strategies to stabilize the GAN training and to overcome the notorious mode collapse problem and ensure a consistent output. DeepUrbanMapper has been extensively evaluated quantitatively and qualitatively for its performance. The evaluation shows that the new framework significantly outperforms the existing methods in terms of visual realism of the translated image and its ability to retain detailed features of the input satellite image which opens up many possibilities urban planning, car navigation systems, GIS.

Attributes Classification for Elaborating the Information of Digital and Imaging Mapping

The rapid development of information has made it possible for everyone to obtain the latest information, complete and accurate, in real time, anytime, anywhere, all over the world. Any information is fine catching up by updated with the latest news and even detailed information on local conditions. With the same analogy, detailed information regarding land utilization, land containing, landscape provision, earth surface contour, etc., are required to inform and elaborate any appropriate decision needed. The Geographic information systems (GIS) is a recent technology commonly used by research in earth science to facilitate many layered detail information by one way to get up-to-date, detailed information. In this research, the GIS utilizes several types of imaging data such remote sensing images and digitize images. As the name suggests, this system captures detailed geographic information about a location or region. By inputting classified images of remote sensing results into a GIS database at regular intervals (adjusted as necessary, such as every year, every two years, every three years, etc.), the number of information sources that can be obtained increases. There are several reasons for that. First, remote sensing images are images that cover the entire surface of the Earth. Next, remote sensing images are images that contain information about the state of the earth's surface. Third, a variety of information can be obtained by performing appropriate image processing. Furthermore, this research could be elaborated by implementing an artificial intelligent algorithm to create a robust outcome.

Deep Learning Models Used for Crop Analysis: A Review

The development and improvement to map agricultural land cover are major challenges for researchers. For the sustainable development of agronomics spatial information about agricultural practices plays a vital role. Remote sensing satellite imagery is a valuable aid in providing and understanding this spatial distribution of agricultural practices. The aim of this paper is to provide a better understanding of capabilities of satellite images for agricultural land cover mapping through the use of deep learning techniques. The global coverage , rich spectral and spatial information and repetitive nature of remote sensing(RS) data have made them effective tools for mapping crop extents and yield prediction. This paper explores wide ranging review of research papers and articles on deep learning algorithms for image processing and predictions in the field of agriculture. The DL algorithms has attained remarkable success in different fields of RS and its use in crop monitoring. This review systematically identified 40 research papers from peer reviewed scientific publications related to sensors, platforms, input features, training data, spatial distribution of study sites. This article provides a concise summary of major DL algorithms, including concepts, limitations, implementation, to help researchers in agriculture to gain a holistic picture of major DL techniques quickly.

Bilateral enhancement network with signal-to-noise ratio fusion for lightweight generalizable low-light image enhancement

Low-light image enhancement aims to enhance the visibility and contrast of low-light images while eliminating complex degradation issues such as noise, artifacts, and color distortions. Most existing low-light image enhancement methods either focus on quality while neglecting computational efficiency or have limited learning and generalization capabilities. To address these issues, we propose a Bilateral Enhancement Network with signal-to-noise ratio fusion, called BiEnNet, for lightweight and generalizable low-light image enhancement. Specifically, we design a lightweight Bilateral enhancement module with SNR (Signal-to-Noise Ratio) Fusion (BSF), which serves the SNR map of the input low-light image as the interpolation weights to dynamically fuse global brightness features and local detail features extracted from a bilateral network and achieve differentiated enhancement across different regions. To improve the network’s generalization ability, we propose a Luminance Normalization (LNM) module for preprocessing and a Dual-Exposure Processing (DEP) module for post-processing. LNM divides the channels of input features into luminance-related channels and luminance independent channels, and reduces the inconsistency of the degradation distribution of input low-light images by only normalizing the luminance-related channels. DEP learns overexposure and underexposure corrections simultaneously by employing the ReLU activation function, inverting operation, and residual network, which can improve the robustness of enhancement effects under different exposure conditions while reducing network parameters. Experiments on the LOL-V1 dataset shows BiEnNet significantly increased PSNR by 8.6\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\%$$\\end{document} and SSIM by 3.6\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\%$$\\end{document} compared to FLW-Net, reduced parameters by 98.78\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\%$$\\end{document}, and improved computational speed by 52.64\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\%$$\\end{document} compared to the classical KIND.

Advancing Sugarcane Farm Management Through NDVI-Based Color Mapping and Drone Imaging

Precision agriculture has emerged as a key strategy for boosting crop productivity and optimizing resource use. This study leverages advanced imaging and machine learning to enhance the management of sugarcane farms. Using drones, high-resolution RGB images of sugarcane fields are captured and transformed into multispectral images through a Generative Adversarial Network (GAN), revealing critical spectral data for plant health assessment. The Normalized Difference Vegetation Index (NDVI) is derived from these multispectral images and serves as a vital measure of vegetation health. This NDVI data, combined with farmer-reported yield information, creates a comprehensive dataset linking NDVI values to actual crop yields. To predict sugarcane yield from NDVI values, we trained a feedforward neural network on this integrated dataset. The proposed method not only enhances prediction accuracy but also provides valuable insights into the connection between NDVI metrics and crop performance. The model was validated using individual field images, enabling precise yield predictions for different field sections. This study highlights the effectiveness of integrating drone imagery, machine learning, and remote sensing in precision agriculture. The combination of NDVI data with yield information provides a robust tool for optimizing sugarcane production, improving farm management decisions, and advancing agricultural sustainability.

State-to-State Spin-Orbit Changing Collision Dynamics of Vibrationally Excited NO at Collision Energies from 1.4 eV to the Cold Regime.

State-to-state spin-orbit changing collisions of vibrationally excited nitric oxide (NO) with argon (Ar) were studied across a wide collision energy range from 3.5 to 11,200 cm-1 (0.43 meV to 1.4 eV) using two molecular beam geometries. Stimulated emission pumping (SEP) for precise initial state preparation and velocity map imaging (VMI) for detailed scattering image capture were employed. These methods enable the study of quantum-state-resolved differential cross sections (DCSs) and provide comprehensive insight into the collision dynamics over both quantum and classical regimes. Theoretical predictions using quantum mechanical close-coupling (QMCC) calculations based on high-level coupled cluster (CCSD(T)) and multireference configuration interaction (MRCI) potential energy surfaces (PESs) are compared with experimental results enabling the testing of both repulsive and attractive parts of the PESs. This study highlights the challenges in accurately modeling spin-orbit changing collisions and underscores the importance of precise experimental data for validating theoretical models, thereby advancing our understanding of nonadiabatic collision dynamics.

Impact of uterine displacement on T2-weighted image quality in the female pelvic MRI.

Despite the importance of T2-weighted image in clinical practice, artifacts can significantly degrade image quality and affect diagnosis. This study quantitatively analyzed uterine displacement and surveyed the relationship between the image quality of fast-spin-echo-T2-weighted image of the female pelvis and quantitative value of uterine displacement. Overall, 147 women (mean age, 46.0 ± 12.8years; age range, 22-84years) who had undergone pelvic magnetic resonance imaging examination using a 3T- magnetic resonance imaging scanner were included. Two radiologists performed a visual assessment of the fast-spin-echo-T2-weighted image in the sagittal plane in terms of ghosts and motion blur, and classified the image quality into the following three groups: poor, moderate, and excellent. Uterine displacement on half-Fourier acquisition single-shot turbo spin-echo-cine images was calculated, and the maximum amplitude of uterine displacement and summation of uterine displacement were calculated from the displacement map images. The Kruskal-Wallis and Steel-Dwass tests were performed to compare the maximum amplitude of uterine displacement and summation of uterine displacement among the three groups. Poor, moderate, and excellent image qualities were observed in 48, 71, and 28 patients, respectively. The quality of fast-spin-echo-T2-weighted images degraded statistically significantly with P < 0.01 as the maximum amplitude of uterine displacement increased. The summation of uterine displacement in the poor and moderate groups had greater statistical significance with P < 0.01 than that in the excellent group.

Development of a nomogram based on whole-tumor multiparametric MRI histogram analysis to predict deep myometrial invasion in stage I endometrioid endometrial carcinoma preoperatively.

The depth of myometrial invasion determines whether International Federation of Gynecology and Obstetrics stage I endometrioid endometrial carcinoma (EEC) patients undergo lymph node dissection. However, subjective evaluation results relying on magnetic resonance imaging (MRI) are not always satisfactory. To develop a nomogram based on whole-volume tumor MRI histogram parameters to preoperatively predict deep myometrial invasion (DMI) in patients with stage I EEC. This retrospective analysis included 131 EEC patients and a training/validation cohort of 92/39 patients at a 7:3 ratio. The histogram parameters were obtained from multiple sequences (ADC mapping and T2-weighted imaging) within volumes of interest. Univariate analysis, least absolute shrinkage and selection operator (LASSO) regression, and multivariate logistic regression were used for feature selection. The performance of clinical model, histogram model, and histogram nomogram was evaluated by calculating the area under the receiver operating characteristic curve (AUC). Age and two morphological features (maximum anteroposterior tumor diameter on sagittal T2-weighted images [APsag] and the tumor area ratio [TAR]) were selected to construct the clinical model. Five histogram parameters were selected for the creation of the histogram model. The nomogram, which combines the histogram parameters, age, APsag, and TAR, achieved the highest AUCs in both the training and validation cohorts (nomogram vs. histogram vs. clinical model: 0.973 vs. 0.871 vs. 0.934 [training] and 0.972 vs. 0.870 vs. 0.928 [validation]). The MR histogram nomogram can help predict the DMI of patients with stage I EEC preoperatively, assisting physicians in the development of personalized treatment strategies.

Sustainable dual-cathode photoelectro-Fenton-like system at a wide pH range for rapid degradation of emerging pollutants

In-situ Fenton like process was highly promising and energy-efficient for treating emerging organic contaminants. However, the challenge in the production and activation of H2O2 faces its practical application. Herein, a photoelectro-Fenton-like (PEF) system was successfully fabricated with BiVO4 thin-film, graphite (C) and PhC2Cu as photoanode, cathode and photocathode, respectively. The ternary system was applied for visible-light degradation of Tetracycline (TC) at a low potential of 0.5 V. H2O2 was generated at the graphite cathode and then activated to produce •OH by Cu(I) from PhC2Cu photocathode. The photo-generated electrons accelerated the Cu(I)/Cu(II) cycle to achieve sustainable PEF reactions. After 150 min of illumination, 97 % TC was degraded in BiVO4|C|PCu system, superior to BiVO4|C and BiVO4|PCu systems. The outstanding stability and reusability of this PEF system was proved by 10 cycles. Moreover, the vulnerable atomic sites of TC molecule were predicated by the Fukui function. Both inhibition zone tests of bacteria and wheat seeds cultivation experiments proved that the constructed PEF system displayed outstanding performance in hampering TC eco-toxicity. For example, the Fv/Fm fluorescence image map demonstrated that the constructed PEF system not only degraded organic pollutants but also lowered their toxicity. Finally, the mechanism of in-situ generation and activation of H2O2 was proposed based on the results of quenching experiments, PL spectra and ESR technique. Overall, this study provides a promising pathway to remediate environmental pollution through constructing a PEF system based on dual cathodes.

Imaging geometric simulation and vibration influence analysis of rotary scanning remote sensing satellites

In response to the lack of in-orbit image and vibration impact measurement methods for new rotary scanning satellites, this paper proposed a satellite imaging geometric simulation method based on 3D rendering engine. Simulated images that reflect strict object-image mapping and real-time dynamic image motion compensation process of TDI cameras are generated. On this basis, the consistency of peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) in measuring vibration influence is analyzed, and a theoretical approximate solution for the sensitive frequency of sinusoidal vibration is obtained, which is inversely proportional to the number of rows of the image to be evaluated and the camera integration time. Finally, the correctness of the image motion compensation model is demonstrated through simulations under different working conditions, and the relative deviation between the experimental values of sensitive frequencies and theoretical results is within −15% to 20%.

A Novel Adaptive Sand Cat Swarm Optimization Algorithm for Feature Selection and Global Optimization.

Feature selection (FS) constitutes a critical stage within the realms of machine learning and data mining, with the objective of eliminating irrelevant features while guaranteeing model accuracy. Nevertheless, in datasets featuring a multitude of features, choosing the optimal feature poses a significant challenge. This study presents an enhanced Sand Cat Swarm Optimization algorithm (MSCSO) to improve the feature selection process, augmenting the algorithm's global search capacity and convergence rate via multiple innovative strategies. Specifically, this study devised logistic chaotic mapping and lens imaging reverse learning approaches for population initialization to enhance population diversity; balanced global exploration and local development capabilities through nonlinear parameter processing; and introduced a Weibull flight strategy and triangular parade strategy to optimize individual position updates. Additionally, the Gaussian-Cauchy mutation strategy was employed to improve the algorithm's ability to overcome local optima. The experimental results demonstrate that MSCSO performs well on 65.2% of the test functions in the CEC2005 benchmark test; on the 15 datasets of UCI, MSCSO achieved the best average fitness in 93.3% of the datasets and achieved the fewest feature selections in 86.7% of the datasets while attaining the best average accuracy across 100% of the datasets, significantly outperforming other comparative algorithms.

Geographically-Informed Modeling and Analysis of Platform Attitude Jitter in GF-7 Sub-Meter Stereo Mapping Satellite

The GF-7 satellite, China’s inaugural sub-meter-level stereoscopic mapping satellite, has been deployed for a wide range of applications, including natural resource investigation, environmental monitoring, fundamental surveying, and the development of global geospatial information resources. The satellite’s stable platform and reliable imaging systems are crucial for achieving high-quality imaging and precise attitude measurements. However, the satellite’s operation is affected by both internal and external factors, which induce vibrations in the satellite platform, thereby affecting image quality and mapping accuracy. To address this challenge, this paper proposes a novel method for constructing a satellite platform vibration model based on geographic location information. The model is developed by integrating composite data from star sensors and gyroscopes (gyro) with subsatellite point location data. The experimental methodology involves the composite processing of gyro data and star sensor optical axis angles, integration of the processed data through time-matching and normalization, and denoising of the integrated data, followed by trigonometric fitting to capture the periodic characteristics of platform vibrations. The positions of the satellite substellar points are determined from the satellite orbit data. A rigorous geometric imaging model is then used to construct a vibration model with geographic location correlation in combination with the satellite subsatellite point positions. The experimental results demonstrate the following: (1) Over the same temporal range, there is a significant convergence in the waveform similarities between the gyro data and the star sensor optical axis angles, indicating a strong correlation in the jitter information; (2) The platform vibration exhibits a robust correlation with the satellite’s geographic location along its orbit. Specifically, the model reveals that the GF-7 satellite experiences the maximum vibration amplitude between 5° S and 20° S latitude during its ascending phase, and the minimum vibration amplitude between 5° N and 20° N latitude during the descending phase. The model established in this study offers theoretical support for optimizing satellite attitude and mitigating platform vibrations.

Myocardial involvement in end-stage renal disease patients with anemia as assessed by cardiovascular magnetic resonance native T1 mapping: An observational study.

Cardiovascular disease has become to the main cause of death in the patients with end-stage renal disease (ESRD), and anemia is associated with increased cardiovascular morbidity and mortality in these patients. This study aimed to explore the impact of anemia on myocardial fibrosis using T1 mapping technique in patients with ESRD. A total of 128 subjects including 98 ESRD patients (65 with anemia, 33 without anemia) and 30 normal controls were enrolled. All subjects were underwent cardiovascular magnetic resonance to obtain cardiac cine and T1 mapping images. As potential markers of fibrosis, native T1 values and global longitudinal strain derived by feature-tracking technique were compared. Differences between 3 groups were analyzed using one-way analysis of variance. Associations between variables were assessed by Pearson and Spearman correlation coefficient appropriately. An independent association was identified by the multiple stepwise linear regression analysis. Intraclass correlation was applied to assess observer variability. In all ESRD patients, native T1 values were significantly longer than those of normal controls (global T1, 1357 ± 42 ms vs 1275 ± 48 ms, P < .001). Global T1 value in ESRD patients with anemia was significantly higher (1375 ± 36 ms) compared to that in ESRD patients without anemia (1322 ± 25 ms) and normal controls (1275 ± 48 ms), respectively (all P < .001). Global T1 correlated with hemoglobin negatively (R= -0.499, P < .001). Multiple stepwise linear regression analysis presented the anemia is independently associated with global T1 (R = 0.607, P < .001). Global longitudinal strain was remarkably reduced in ESRD patients with anemia in comparison to those without anemia (P < .001). Diffuse myocardial fibrosis could be detected by native T1 mapping in ESRD patients with long-term anemia. Anemia is an important factor in myocardial fibrosis in ESRD patients, and the evaluation of myocardial involvement is worth considering for clinical management.

Controllability and diffeomorphism groups on manifolds with boundary

Abstract In this article we consider diffeomorphism groups of manifolds with smooth boundary. We show that the diffeomorphism groups of the manifold and its boundary fit into a short exact sequence which admits local sections. In other words, they form an infinite-dimensional fibre bundle. Manifolds with boundary are of interest in numerical analysis and with a view towards applications in machine learning we establish controllability results for families of vector fields. This generalises older results due to Agrachev and Caponigro in the boundary-less case. Our results show in particular that the diffeomorphism group of a manifold with smooth boundary is generated by the image of the exponential map.

Relationship between high-level color features and temperature mapping of magnesium alloy surface images based on the K-nearest neighbor algorithm

Numerous deformation and processing application scenarios of magnesium alloy sheets require high precision and efficiency in temperature detection. Traditional contact temperature measurement and infrared thermal imaging temperature measurement methods suffer from low accuracy and susceptibility to environmental influences. To this end, a real-time sensing method for determining the average surface temperature of magnesium alloys using visible light images is proposed in this paper. Firstly, real-time surface images of 3 mm thick magnesium alloy sheets were captured as the temperature decreased from 450 °C to 300 °C under varying light intensities. Then, calculate the high-level color features based on the grayscale frequency of the Red Green Blue color space of the image, and select the high-level color features with a strong correlation with temperature by Fisher criterion. Finally, a mapping relationship between the high-level color features of magnesium alloy images and temperature was established using the K-nearest neighbor algorithm. The results show that images in RAW format have a wider grayscale range compared to JPG, which is helpful in determining the mapping relationship between image color features and temperature. After extracting and selecting the high-level color features that have a strong correlation with temperature, the dimensions of input features are reduced. Using the K-nearest neighbor algorithm can fit the complex nonlinear relationship between high-level color features and temperature more accurately.