Multispectral Measurements Research Articles

Multispectral and multiangle measurements of acoustic seabed backscatter acquired with a tilted calibrated echosounder.

A multispectral and multiangle analysis of seabed backscatter intensity has been conducted using data from a calibrated single-beam echosounder (SBES) with five frequency channels deployed over four homogeneous areas with different sediment types in the Bay of Brest (France). The SBES transducers were tilted at incidence angles from 0° to 70° to record the seafloor backscatter angular response at discrete frequencies ranging from 35 to 450 kHz. The recorded backscatter levels were analyzed for their angular dependence (average backscatter strength versus frequency and angle) as well as for their sample statistical distribution. The angle and frequency dependence of the seafloor backscatter obtained using a calibrated SBES can potentially be used to calibrate multibeam systems, and it can also help in elucidating the physical processes of backscatter controlled by the interaction between the acoustic wave characteristics and the sediment properties. Backscatter measurements for each area showed a consistent frequency dependence with little variation between the four sediment types.

Polarimetric sensitivity of light-absorbing carbonaceous aerosols over ocean: A theoretical assessment

Visible-light-absorbing carbonaceous aerosols within the boundary layer affect the radiance and polarization states of the radiation at the top of the atmosphere. Remote sensing from suborbital and satellite-based platforms utilizes these radiance and polarization signals to retrieve the key properties of these aerosols. Recent retrieval algorithms have shown a progressive trend toward including multi-angular and multi-spectral polarimetric measurements to produce better retrieval accuracy in comparison to those using measurements based on a single viewing angle. Here, we perform a theoretical investigation of the top of atmosphere (TOA) radiance-related reflectance factor (bidirectional reflectance factor (BRF)) and the two types of polarimetry-related factors (polarized bidirectional reflectance factor (pBRF) and the degree of linear polarization (DoLP)) for different types of atmospheric light-absorbing carbonaceous aerosols as a function of particle size distribution. We selected three polarimetric bands corresponding to those utilized by NASA's Airborne Multiangle SpectroPolarimetric Imager (AirMSPI)—near-UV (470 nm), visible (660 nm), and near-infrared (865 nm)—for our simulations which were performed over ocean surface using the successive order of scattering (SOS) algorithm coupled to a Lorenz-Mie aerosol optics model. The analysis of particle phase matrix elements indicates a close relationship between the angular dependencies of DoLP and associated phase matrix components at the shortest polarimetric band (470 nm). Using Jacobian analysis, we find that the radiance- and polarimetry-related reflectance factors of weakly light-absorbing aerosols, such as brown carbon, are more sensitive to changes in particle size and imaginary refractive index in comparison with those of black carbon, which is strongly light-absorbing. Our results suggest that the DoLP data could be used by future retrieval algorithms for reliably estimating microphysical properties of absorbing carbonaceous aerosols with imaginary refractive index less than 0.4.

Predicting forest fire using multispectral satellite measurements in Nepal

The present study examined the geospatial datasets concerning fire points, topography, vegetation types, tree cover percent, ecological zone, and climate to understand the prospects of forest fire in Nepal. The GIS-based modeling approaches in association with multispectral satellite-derived data were used for future fire prediction in villages. The analysis highlights the vegetation types such as Tropical lowland, broadleaved, evergreen forests (<1000 m) retain the highest fire risk. The fire intensity was found higher in the area of high tree cover density, lower elevated and lower slopes. The highest number of fires was found in the year 2016 and 91% of the forest fire occurs in March, April, and May i.e., hot and dry pre-monsoon season. The statistical analysis such as Crammer's V coefficient (CVC) significantly highlights a strong association of forest fire with precipitation (0.67) and temperature (0.41). It was found that 302 villages of the total of 3967 villages are found to be a potentially high fire risk in present scenario and 20 villages of the potential fire risk villages are predicted as critically fire risk (greater than 80%) in future. The eventual impact of such fire affects human health, life and property. The study offers the fire prone areas in administrative and ecological levels which is useful for making policy, management, resource allocation and forest fire mitigation.

Compressive spectral image fusion via a single aperture high throughput imaging system

Spectral image fusion techniques combine the detailed spatial information of a multispectral (MS) image and the rich spectral information of a hyperspectral (HS) image into a high-spatial and high-spectral resolution image. Due to the data deluge entailed by such images, new imaging modalities have exploited their intrinsic correlations in such a way that, a computational algorithm can fuse them from few multiplexed linear projections. The latter has been coined compressive spectral image fusion. State-of-the-art research work have focused mainly on the algorithmic part, simulating instrumentation characteristics and assuming independently registered sensors to conduct compressed MS and HS imaging. In this manuscript, we report on the construction of a unified computational imaging framework that includes a proof-of-concept optical testbed to simultaneously acquire MS and HS compressed projections, and an alternating direction method of multipliers algorithm to reconstruct high-spatial and high-spectral resolution images from the fused compressed measurements. The testbed employs a digital micro-mirror device (DMD) to encode and split the input light towards two compressive imaging arms, which collect MS and HS measurements, respectively. This strategy entails full light throughput sensing since no light is thrown away by the coding process. Further, different resolutions can be dynamically tested by binning the DMD and sensors pixels. Real spectral responses and optical characteristics of the employed equipment are obtained through a per-pixel point spread function calibration approach to enable accurate compressed image fusion performance. The proposed framework is demonstrated through real experiments within the visible spectral range using as few as 5% of the data.

Measuring Tumor Microenvironment pH During Radiotherapy Using a Novel Cerenkov Emission Multispectral Optical Probe Based on Silicon Photomultipliers

Cerenkov Emission (CE) multispectral analysis with silicon photomultiplier (SiPM)-based optical probes is a promising tool for online tumor microenvironment interrogation and targeting during radiotherapy delivery. With the extreme sensitivity of SiPMs, deep tissue multispectral CE measurements can be realized in a clinical setting. In this work, we utilize our Cerenkov Emission Multi-spectral Imaging (CMSI) prototype probe to interrogate the spectral components of the CE signal generated during external beam radiotherapy. Our results demonstrated that CMSI enables effective probing of in vitro quantitative changes in the pH of cell media to monitor cancer cell proliferation after various treatment pathways and differentiate between varying treatment resistance cell lines. In addition, the feasibility of using the CMSI probe in vivo was also successfully demonstrated by measuring tumor pH during a pilot mouse study.

UAV-based high-throughput phenotyping to increase prediction and selection accuracy in maize varieties under artificial MSV inoculation

The use of unmanned aerial vehicles’ (UAV) remotely sensed data in crop evaluation is revolutionizing the field of plant phenotyping. This study was conducted to (1) develop protocol to predict maize streak virus (MSV) and grain yield using UAV-derived multispectral data; and (2) identify the suitable predictor variables and ideal phenological stages for MSV and grain yield prediction. Twenty-five maize varieties were evaluated under artificial MSV inoculation. Manual scoring and multispectral imaging measurements were performed at mid-vegetative, flowering and mid-grain filling stages. UAV-derived data were acquired in the multispectral bands of Green (0.53–0.57 μm), Red (0.64–0.68 μm), Red-edge (0.73–0.74 μm) and Near-Infrared (0.77–0.81 μm). Eight vegetation indices were determined: NDVI (normalized difference vegetation index), NDVIred-edge, GNDVI (green normalized difference vegetation index), SR (simple ratio), CIgreen (green chlorophyll index), CIred-edge (red-edge chlorophyll index), SAVI (soil-adjusted vegetation index) and OSAVI (optimized SAVI). Finally, predictions of MSV and grain yield were performed with 36 models using multiple regression, decision trees and linear regression. Frequently selected variables for MSV prediction were Green band at vegetative (61.5%), Red band at vegetative (68.4%) and flowering (80.4%), and GNDVI at mid-vegetative (88.7%). The best MSV predictors were GNDVI (r = 0.84; RMSE = 0.85), CIgreen (r = 0.83; RMSE = 0.86) and Red band (r = 0.77; RMSE = 0.99) measured at mid-vegetative stage. Six out of 36 models were selected as ideal for predicting maize grain yield: RF-REF-NIRF (r = 0.69; RMSE = 0.65); NDVIREG-GNDVIG (r = 0.74; RMSE = 0.56); RV-NIRV (r = 0.84; RMSE = 0.37); and the tree with the largest correlations are RV-NIRV-RF (r = 0.86; RMSE = 0.32); GNDVIV-OSAVIV (r = 0.84; RMSE = 0.36); GV-RV-NIRV (r = 0.84; RMSE = 0.35); the last two of which were at mid-vegetative stage. We conclude that UAV-based multispectral remote sensing is a reliable tool for phenotyping MSV disease and grain yield prediction, and mid-vegetative appear to be the most ideal phenological stage for MSV and grain yield prediction.

Learned spectral decoloring enables photoacoustic oximetry

The ability of photoacoustic imaging to measure functional tissue properties, such as blood oxygenation sO_2, enables a wide variety of possible applications. sO_2 can be computed from the ratio of oxyhemoglobin HbO_2 and deoxyhemoglobin Hb, which can be distuinguished by multispectral photoacoustic imaging due to their distinct wavelength-dependent absorption. However, current methods for estimating sO_2 yield inaccurate results in realistic settings, due to the unknown and wavelength-dependent influence of the light fluence on the signal. In this work, we propose learned spectral decoloring to enable blood oxygenation measurements to be inferred from multispectral photoacoustic imaging. The method computes sO_2 pixel-wise, directly from initial pressure spectra S_{text {p}_0}(lambda , mathbf {x}), which represent initial pressure values at a fixed spatial location mathbf {x} over all recorded wavelengths lambda. The method is compared to linear unmixing approaches, as well as pO_2 and blood gas analysis reference measurements. Experimental results suggest that the proposed method is able to obtain sO_2 estimates from multispectral photoacoustic measurements in silico, in vitro, and in vivo.

Error analysis and reflection correction for radiation temperature measurements at high background temperatures

Advances in science and technology have increased the demand for greater temperature measurement accuracy. Radiation thermometry is a mature technology that is applicable to various industrial fields. In special environments (e.g. a high ambient temperature), background radiation reflected by the target surface is superimposed with the radiation from the target itself, which affects the measured radiation temperature. In this study, different radiation thermometry methods were evaluated. The errors of single-spectral, colorimetric, and multispectral temperature measurements of targets in a high-temperature environment have been discussed. Based on this, we propose a multi-wavelength temperature measurement method with reflection correction to reduce the impact of high-temperature environment. Experiments show that this method can reduce the uncertainty in the multispectral temperature measurement from 4.16% to 0.26% under the experimental conditions.

The SARSense Campaign: Air- and Space-Borne C- and L-Band SAR for the Analysis of Soil and Plant Parameters in Agriculture

With the upcoming L-band Synthetic Aperture Radar (SAR) satellite mission Radar Observing System for Europe L-band SAR (ROSE-L) and its integration into existing C-band satellite missions such as Sentinel-1, multi-frequency SAR observations with high temporal and spatial resolution will become available. The SARSense campaign was conducted between June and August 2019 to investigate the potential for estimating soil and plant parameters at the agricultural test site in Selhausen (Germany). It included C- and L-band air- and space-borne observations accompanied by extensive in situ soil and plant sampling as well as unmanned aerial system (UAS) based multispectral and thermal infrared measurements. In this regard, we introduce a new publicly available SAR data set and present the first analysis of C- and L-band co- and cross-polarized backscattering signals regarding their sensitivity to soil and plant parameters. Results indicate that a multi-frequency approach is relevant to disentangle soil and plant contributions to the SAR signal and to identify specific scattering mechanisms associated with the characteristics of different crop type, especially for root crops and cereals.

Atmospheric correction over the ocean for hyperspectral radiometers using multi-angle polarimetric retrievals.

We developed a fast and accurate polynomial based atmospheric correction (POLYAC) algorithm for hyperspectral radiometric measurements, which parameterizes the atmospheric path radiances using aerosol properties retrieved from co-located multi-wavelength multi-angle polarimeter (MAP) measurements. This algorithm has been applied to co-located spectrometer for planetary exploration (SPEX) airborne and research scanning polarimeter (RSP) measurements, where SPEX airborne was used as a proxy of hyperspectral radiometers, and RSP as the MAP. The hyperspectral remote sensing reflectance obtained from POLYAC is accurate when compared to Aerosol Robotic Network (AERONET), and Visible Infrared Imaging Radiometer Suite (VIIRS) ocean color products. POLYAC provides a robust alternative atmospheric correction algorithm for hyperspectral or multi-spectral radiometric measurements for scenes involving coastal oceans and/or absorbing aerosols, where traditional atmospheric correction algorithms are less reliable.

Multispectral radiometric temperature measurement algorithm for turbine blades based on moving narrow-band spectral windows.

This paper addresses the problem of inaccurate emissivity presets for multispectral temperature measurements of aero-engine turbine blades and proposes a narrow-band spectral window moving temperature inversion algorithm that does not rely on an assumed emissivity model. As the emissivity of the measured object changes slowly over the narrow spectral window, the temperature corresponding to the normalized spectral radiation intensity for each window in the set temperature range is calculated using the Mahalanobis distance coefficient. The temperature error is less than 1.33% relative to thermocouple measurements when using this algorithm to perform temperature inversion on the experimental spectrum curves for different types of alloy samples. Furthermore, a two-dimensional spectral temperature field measurement platform was built, and the surface temperature fields of alloy samples were reconstructed using the narrow-band spectral window moving algorithm. The proposed algorithm is shown to provide high-precision inversion of the temperature field without presetting the emissivity model, which gives a new processing concept for the application of infrared spectral temperature measurements.

Four-Channel Differential Lock-in Amplifiers With Autobalancing Network for Stimulated Raman Spectroscopy

We introduce a multi-channel integrated circuit for fast stimulated Raman scattering (SRS) microscopy on multiple simultaneous frequencies at a frame rate higher than 1 frame/s. The chip is a four-channel differential readout system, based on the lock-in technique. It is able to measure down to 10 ppm SRS signal, over a wide range of input optical powers (50-600 μW per channel), with a pixel dwell time of only 30 μs. Each acquisition channel includes two low-noise preamplifiers, two variable-gain amplifiers, a fully differential voltage subtractor, and a lock-in demodulator. The differential readout electronics rejects the power fluctuations of the laser, and it is automatically balanced by an analog feedback loop over ±30% input power mismatches. Due to the autobalancing network, the pixel dwell time is reduced by a factor up to 225 with a settling time of only 10 μs. The chip is fabricated in AMS 0.35 μm CMOS technology, and it is included in a combined electronics and optical system. Both single-pixel spectral measurements and multi-spectral imaging measurements are presented to validate the full SRS microscope.

Hyperspectral Satellite Radiance Atmospheric Profile Information Content and Its Dependence on Spectrometer Technology

Satellite vertical atmospheric sounding was initiated more than 50 years ago and has evolved to provide the most critical component of today's global observation system. However, the operational use of today's polar orbiting satellite hyperspectral infrared (IR) observations in numerical weather prediction (NWP) has been limited to a small fraction of the radiance information being provided. On the other hand, research systems are in operation that combines high vertical resolution polar hyperspectral radiance measurements with high spatial and time resolution geostationary multispectral radiance measurements that demonstrate the promise of future geo-hyperspectral sounding observations to significantly improve the forecast location and warning time for the development of localized tornadic storms. This article has a twofold objective: 1) to demonstrate that there is much more information available in current IR sounding data, than is being used to benefit the current NWP operation and 2) to illustrate the importance of the spectrometer technology (i.e., Fourier transform vs. dispersive grating) used for achieving the vertical profile resolution required to improve both extended range and localized severe weather forecasts. These objectives are achieved by performing both theoretical physics-based radiance information content (IC) studies and empirical analyses of current hyperspectral radiance measurements. The IC studies clearly demonstrate the unique importance of longwave IR (9−15 μm) radiance observations. The empirical studies demonstrate the importance of using Fourier transform spectrometers for providing the high spectral fidelity needed to resolve the small-scale vertical features in atmospheric temperature and moisture profiles, which impact weather forecast accuracy.

In-situ measurement of temperature and potassium concentration during the combustion of biomass pellets based on the emission spectrum

Here, a calibration method using the K concentration, temperature and the characteristic spectral line radiation intensity in flame is described. In addition, a radiation function model between these three variables, Iλ766T,CK=0.821·Ib(λ766,T)·CK, was established. A Y-type optical fiber was used to connect the spectral signals from the same direction to two spectrometers. The continuous spectra in the 500–800 nm band were used for multi-spectral temperature measurements. The combination of the measured temperature and the K characteristic spectral line radiation intensity at 766 nm was used for measurements of K concentration. The temperature and equivalent ratio of the heating atmosphere were 1300 K (±20 K) and 0.786, respectively. Compared with rice straw pellet, pine wood pellet had a lower K concentration and higher combustion temperature during the entire combustion process. The peak K release concentrations of the pine wood pellet and the rice straw pellet during the volatile combustion stage were 2.09 ppm and 10.03 ppm, respectively, and the peak flame temperatures were 1774 K and 1756 K, respectively. In the char combustion stage, the peak K release concentrations were 27.83 ppm and 110.43 ppm, respectively, and the peak surface temperatures were 1287 K and 1143 K, respectively. The established function model for measuring K concentration is also applicable to the combustion diagnosis of other K-rich fuels.

Multifunctional optical fiber sensor for simultaneous measurement of temperature and salinity.

A multifunctional optical fiber sensor fabricated by asymmetric offset splicing is proposed in this Letter. The light is divided into several parts at the offset interface, among which the transmitted light forms the Mach-Zehnder interference (MZI) spectrum while the reflected light forms the Fabry-Perot interference (FPI) spectrum. The online monitoring system is built to create a better light distribution at the offset interface. Theoretical analysis and experimental verification are carried out. The results of the experiment show that the proposed sensor has good characteristics of salinity and temperature, and the salinity sensitivity is as high as -2.4473nm/‰ in the range of 20-40‰; the temperature sensitivity is better than 2.17 nm/°C in the range of 28-48 °C. The two interferometers involved have different responses to temperature and salinity, contributing to the effective elimination of cross-sensitivity. The proposed optical fiber sensor has the benefits of compact size, high sensitivity, and multispectral measurement function.

Drone data reveal heterogeneity in tundra greenness and phenology not captured by satellites

Data across scales are required to monitor ecosystem responses to rapid warming in the Arctic and to interpret tundra greening trends. Here, we tested the correspondence among satellite- and drone-derived seasonal change in tundra greenness to identify optimal spatial scales for vegetation monitoring on Qikiqtaruk—Herschel Island in the Yukon Territory, Canada. We combined time-series of the Normalised Difference Vegetation Index (NDVI) from multispectral drone imagery and satellite data (Sentinel-2, Landsat 8 and MODIS) with ground-based observations for two growing seasons (2016 and 2017). We found high cross-season correspondence in plot mean greenness (drone-satellite Spearman’s ρ 0.67–0.87) and pixel-by-pixel greenness (drone-satellite R 2 0.58–0.69) for eight one-hectare plots, with drones capturing lower NDVI values relative to the satellites. We identified a plateau in the spatial variation of tundra greenness at distances of around half a metre in the plots, suggesting that these grain sizes are optimal for monitoring such variation in the two most common vegetation types on the island. We further observed a notable loss of seasonal variation in the spatial heterogeneity of landscape greenness (46.2%–63.9%) when aggregating from ultra-fine-grain drone pixels (approx. 0.05 m) to the size of medium-grain satellite pixels (10–30 m). Finally, seasonal changes in drone-derived greenness were highly correlated with measurements of leaf-growth in the ground-validation plots (mean Spearman’s ρ 0.70). These findings indicate that multispectral drone measurements can capture temporal plant growth dynamics across tundra landscapes. Overall, our results demonstrate that novel technologies such as drone platforms and compact multispectral sensors allow us to study ecological systems at previously inaccessible scales and fill gaps in our understanding of tundra ecosystem processes. Capturing fine-scale variation across tundra landscapes will improve predictions of the ecological impacts and climate feedbacks of environmental change in the Arctic.

Abstract 15508: Random Forest Classifier-incoporated Intravascular Optical Coherence Tomography-fluorescence Lifetime Imaging (oct-flim) Provides Automated Characterization of Key Biochemical Components of Coronary Atherosclerotic Plaques

Introduction: Fluorescence lifetime imaging (FLIm) allows label-free biochemical characterization of atheroma, however, multispectral FLIm produces massive data throughput making image interpretation problematic. We investigated whether machine learning algorithm can be applied to intravascular OCT-FLIm for automated biochemical characterization of coronary plaques. Methods and Results: We built a fully-integrated, high-speed OCT-FLIm system and a low-profile, dual-modal imaging catheter that can provide high-resolution OCT images and correctly co-registered multispectral fluorescence lifetime (FL) readouts: ch.1 and ch.2 FL, and FL intensity ratio (ch.2/ch.1). Rapid intracoronary imaging (10-20 mm/s, 100 rps) was safely performed in atheromatous pigs. There were significant differences in FL measurements according to plaque types (high-risk vs. fibrotic: p&lt;0.001). Multispectral FL measurements sampled selectively from histologically-proven plaque components (lipid, macrophage, lipid+macrophage, SMC) were analyzed. Each component was distinguishable from one another either by difference of FLs or intensity ratio. Random forest classifier (RFC), trained with component-labeled multispectral FL dataset, accurately classified key biochemical components of atherosclerotic plaques. RFC-determined biochemical characteristics of target plaque were consistent across two repeated imaging data (intraclass correlation, p&lt;0.0001) and corroborated closely with those derived from quantitative immunohistochemistries. Conclusions: Our OCT-FLIm incorporating RFC-based systematic multispectral FL analysis could provide high-resolution plaque imaging with automated biochemical characterization in beating coronary environment. The present imaging strategy enabling comprehensive characterization of multiple components of atherosclerotic plaques will open a new avenue in the field of cardiovascular imaging.

Monitoring formaldehyde in a shock tube with a fast dual-comb spectrometer operating in the spectral range of 1740\u20131790\xa0cm\u20131

A dual-frequency-comb spectrometer based on two quantum-cascade lasers is applied to kinetics studies of formaldehyde (HCHO) in a shock tube. Multispectral absorption measurements are carried out in a broad spectral range of 1740–1790 cm–1 at temperatures of 800–1500 K and pressures of 2–3 bar. The formation of HCHO from thermal decomposition of 1,3,5-trioxane (C3H6O3, 0.9% diluted in argon) and the subsequent oxidation of formaldehyde is monitored with a time resolution of 4 µs. The rate coefficient of the decomposition of C3H6O3 (i.e., HCHO formation) is found to be k1 = 6.0 × 1015 exp(− 205.58 kJ mol−1/RT) s–1. For the oxidation studies, mixtures of 0.36% C3H6O3 and 1% O2 in argon are used. The information of all laser lines, along with the consideration of individual signal variance of each line, is utilized for kinetic and spectral analysis. The experimental kinetic profiles of HCHO are compared with simulations based on the mechanisms of Zhou et al. (Combust Flame, 197:423–438, 2018) and Cai and Pitsch (Combust Flame, 162:1623–1637, 2015).

NASA’s Multi-Angle Imager for Aerosols (MAIA): Applications of NASA Satellite Data to Airborne Particulate Matter Exposure and Human Health

The upcoming NASA Multi-Angle Imager for Aerosols (MAIA) investigation seeks to extend NASA’s capabilities to study the impact of different size and compositional mixtures of particulate matter (PM) on adverse health outcomes. The MAIA satellite instrument, in development at the Jet Propulsion Laboratory and planned for launch in mid-2022, will collect multiangular, multispectral, and polarimetric measurements over a set of globally distributed target areas. Retrievals of aerosol properties will be combined with ground-based monitor data and chemical transport modeling to produce 1-km gridded data products of daily-averaged PM10 and PM2.5 mass, as well as the fractional abundances of sulfate, nitrate, organic carbon, elemental carbon, and dust making up the retrieved PM2.5 mixtures. Epidemiologists on the MAIA Science Team and their collaborators will use these data in studies aimed at associating health risk with particle types. The MAIA Early Adopters Program is designed to engage the wider air quality and public health communities and assist them in using MAIA data products, which will be publicly available, free of charge, from NASA’s Atmospheric Science Data Center. This presentation will cover the plans for operational MAIA data products as well as test data products that will be available prior to launch, and opportunities for Early Adopters to become involved in the program.

Multi-spectral temperature measurement based on adaptive emissivity model under high temperature background

When the object surface temperature in high temperature environment is measured by multi-spectral temperature measurement method, the radiation generated by the reflection of the high temperature background on the surface of the object and the radiation emitted by the object itself are both received by the pyrometer. For one thing, the amount of radiation received by the pyrometer increases. On the other hand, the surface emissivity model may change. So a large temperature measurement error may occur. In order to accurately measure the true surface temperature of an object in a high temperature environment, a multi-spectral temperature measurement method based on adaptive emissivity model is proposed in this paper. First, BP neural network is used to identify the shape of spectral data and obtain the correspondence between spectral data and emissivity model, then output the emissivity model that conforms to the measured target and specific environment. Moreover, the traditional NSGA-II algorithm (Non-dominant sorting genetic algorithm-II) is prone to falling into local optimal solutions and the numerical values obtained are not accurate enough. In this paper, the concepts of clustering algorithm, difference operator and symmetrical solution are introduced in the process of crossover and mutation, and then the new algorithm of INSGA-II (Improved Non-dominant sorting genetic algorithm-II)) is proposed. On the basis of this algorithm, the emissivity model of BP network is combined to solve the target true temperature. The experimental results show that the maximum temperature measurement error can reach 60 K if the influence of high temperature background is ignored. After removing the influence of high temperature background, the non-adaptive emissivity model was used to obtain the maximum temperature error of 30 K, and the fluctuation between adjacent temperature points was large. The multi-spectral temperature measurement method based on adaptive emissivity model proposed in this paper can obtain more accurate temperature measurement results with the maximum error of 8 K.