Multispectral Satellite Images Research Articles

Developing a novel index for detection of optically shallow waters using multispectral satellite imagery and radiative transfer modelling

ABSTRACT Optical remote sensing of water quality has been problematic due to contamination of remote sensing observations by the sea-bottom effect over complex shallow or very clear waters. This has resulted in providing misleading information on the estimation of Water Constituent Concentrations (WCCs) retrieved from satellite images. In this research, we used Radiative Transfer (RT) Modelling to develop a simple and innovative index named the Sea-Bottom Effect Index (SEBEI) to readily determine water pixels contaminated by the sea-bottom effect, (hereafter named Optically Shallow Waters, (OSWs)) from remote sensing observations. To define the SEBEI, we initially assessed the influence of the sea-bottom on simulated water-leaving reflectance (R rs ) using RT Water-Sea-Bottom (WSB) modelling. This evaluation encompassed a range of water depths, seabed albedos, and WCCs (i.e. Chlorophyll-a (Chla) and Suspended Particulate Matter (SPM)). Next, we detected the most sensitive wavelengths (i.e. 750 nm, 810 nm, and 900 nm) to the seabed contribution on the simulated R rs spectra at the water surface level and developed the SEBEI. We validated the accuracy of the proposed SEBEI over a time series of MEdium Resolution Imaging Spectrometer (MERIS) and Ocean and Land Colour Instrument (OLCI) images captured during low and high tidal phases over shallow waters of the Dutch Wadden Sea, the Netherlands. The results indicate that the SEBEI effectively distinguishes between OSWs and Optically Deep Waters (ODWs) using satellite images (with an R2 value of ≥ 0.97 and a Root Mean Square Error (RMSE) of ≤ 0.65). The SEBEI can serve as an intermediate solution for detecting OSWs in multispectral and hyperspectral satellite images worldwide. It eliminates the need for ancillary datasets like bathymetry maps and significantly enhances the reliability of WCC maps produced over complex shallow coastal waters.

Using Multi-Source National Forest Inventory Data for the Prediction of Tree Lists of Individual Stands for Long-Term Simulation

Forest resource maps and small area estimates have been produced by combining national forest inventory (NFI) field plot data, multispectral satellite images and numerical map data. We evaluated k-nearest neighbors (k-NN) method-based predictions of forest variables for pixels in predicting tree lists of individual stands, including tree diameters at breast height and tree heights and then calculated stem volumes and tree species proportions. We compared alternative parameters (k-NN) using k of either 1 or 5 according to preliminary plot-level study and applying either measured trees (1-NN_trees) or mean stand characteristics (k-NN_stand). In the 1-NN_trees method, a tree list was generated based on the measured trees of the NFI plots, whereas in the 1-NN_stand and 5-NN_stand methods, a Weibull-based diameter distribution was recovered from the stand characteristics of the same inventory plots. In both methods, tree lists were predicted for each 16 m × 16 m pixel included in the stand compartment. Both methods performed well and resulted in 8–14% differences in the total volume compared with the field inventory of the 27 stands used for the evaluation. Moreover, the main tree species was correctly predicted for 74% of cases. The RMSE in total volume ranged from 25% (5-NN_stand) to 31% (1-NN_stand), while the smallest RMSEs in volume by tree species were 61% for broadleaves and 65% for pine and spruce using the 5-NN_stand. When comparing input data for a long-term growth simulation, the choice of the method was less influential as the effect of the error in the initial stand characteristics decreased over time during the simulation period. After 30-year simulation of the inventoried stands, the respective RMSEs were 9.4% for total volume and 39%, 50% and 59% for tree species, respectively. The satellite-based data with NFI plots were useful for predicting tree lists for pixels of a stand. However, the accuracy for operational forest management was still questionable. For a larger area’s strategic information, the accuracy is considered adequate.

Using Multi-Source National Forest Inventory Data for the Prediction of Tree Lists of Individual Stands for Long-Term Simulation

Forest resource maps and small area estimates have been produced by combining national forest inventory (NFI) field plot data, multispectral satellite images and numerical map data. We evaluated k-nearest neighbors (k-NN) method-based predictions of forest variables for pixels in predicting tree lists of individual stands, including tree diameters at breast height and tree heights and then calculated stem volumes and tree species proportions. We compared alternative parameters (k-NN) using k of either 1 or 5 according to preliminary plot-level study and applying either measured trees (1-NN_trees) or mean stand characteristics (k-NN_stand). In the 1-NN_trees method, a tree list was generated based on the measured trees of the NFI plots, whereas in the 1-NN_stand and 5-NN_stand methods, a Weibull-based diameter distribution was recovered from the stand characteristics of the same inventory plots. In both methods, tree lists were predicted for each 16 m × 16 m pixel included in the stand compartment. Both methods performed well and resulted in 8–14% differences in the total volume compared with the field inventory of the 27 stands used for the evaluation. Moreover, the main tree species was correctly predicted for 74% of cases. The RMSE in total volume ranged from 25% (5-NN_stand) to 31% (1-NN_stand), while the smallest RMSEs in volume by tree species were 61% for broadleaves and 65% for pine and spruce using the 5-NN_stand. When comparing input data for a long-term growth simulation, the choice of the method was less influential as the effect of the error in the initial stand characteristics decreased over time during the simulation period. After 30-year simulation of the inventoried stands, the respective RMSEs were 9.4% for total volume and 39%, 50% and 59% for tree species, respectively. The satellite-based data with NFI plots were useful for predicting tree lists for pixels of a stand. However, the accuracy for operational forest management was still questionable. For a larger area’s strategic information, the accuracy is considered adequate.

RDC-UNet++: An end-to-end network for multispectral satellite image enhancement

Multi-spectral satellite imagery is an ideal data source for comprehensive, real-time Earth observation (EO) due to its extensive coverage of Earth and regular updates. It has a wide range of applications in environment monitoring, disaster management, urban planning, weather forecasting etc. Yet, the visual aspect of these images and thus the possibility to extract useful information using image processing techniques is often degraded due to fog, rain, dust, cloud, etc. Satellite image enhancement denotes a set of techniques designed to improve the quality of a satellite image such that the result is more useful for image analysis. The image enhancement aims to improve the quality of an image such that the enhanced image is more useful for image analysis than the original image for a particular remote sensing application. This study introduces a multi-spectral satellite image enhancement architecture called Residual Dense Connection-based UNet++ (RDC-UNet++). The unique design can improve multi-spectral images by enhancing their color and texture details. Extensive experimental studies on multi-spectral image datasets containing more than 150 images prove that the proposed architecture performs better than recent state-of-the-art satellite image enhancement algorithms.

Deep learning-based harmonization and super-resolution of Landsat-8 and Sentinel-2 images

Multi-spectral satellite images of the Earth’s surface are used in various applications, from water quality assessment and urban planning to climate monitoring, disaster response, infrastructure oversight, and agricultural surveillance. Many of these applications would benefit from higher spatial and temporal resolution of observations, which could be achieved by combining observations from several sources. This study introduces a deep learning-based pipeline to harmonize the spectral and spatial discrepancies between the Landsat-8 and Sentinel-2 Earth Observation satellites. Through established image quality metrics, we demonstrate a significant enhancement in the spatial resolution of Landsat-8 images. Field observation experiments show that leveraging unified images from both satellites increases the availability of cloud-free images by 21% annually on average in our study area. Additionally, our pipeline enhances the Normalized Difference Vegetation Index (NDVI) correlation between Landsat-8 and Sentinel-2 observations by about 4.9%, offering significant performance gains in a downstream crop segmentation task. Our 100M parameter model, trained on European data, generalizes to most regions with only minor limitations. Furthermore, we show that the pipeline can provide uncertainty estimates for its outputs, which are valuable for decision-making in downstream applications. Our code base is available online: https://github.com/venkatesh-thiru/Deep-Harmonization.

Monitoring saltwater intrusion to estuaries based on UAV and satellite imagery with machine learning models

Saltwater intrusion is a natural mixture process between watershed freshwater and seawater that frequently occurs in estuaries. Station-based monitoring of saltwater intrusion is time-consuming and labor-intensive. To enable quick monitoring of saltwater intrusion, this study developed new remote sensing algorithms for water surface salinity measurement using four decision tree-based machine learning models. These models were built based on simultaneously collected in-situ salinity data from a waterway in the Pearl River Delta and Unmanned Aerial Vehicle (UAV) hyperspectral images. A 10-fold cross-validation was applied to assess the performance of the models, with XGBoost outperforming the other three models (R2=0.93, RMSE = 0.88 psu). Then the developed model was employed for Sentinel-2 multispectral satellite images to invert the estuarine salinity distribution at a larger spatial scale. Results displayed the high performance of the machine learning models proposed in this study for mapping the salinity distribution in river channels, making it an efficient and practical technique for monitoring saltwater intrusion in river channels at a regional scale.

Satellite Image Fusion Airborne LiDAR Point-Clouds-Driven Machine Learning Modeling to Predict the Carbon Stock of Typical Subtropical Plantation in China

In the current context of carbon neutrality, afforestation is an effective means of absorbing carbon dioxide. Stock can be used not only as an economic value index of forest wood resources but also as an important index of biomass and carbon storage estimation in forest emission reduction project evaluation. In this paper, we propose a data-driven machine learning framework and method for predicting plantation stock based on airborne LiDAR + satellite remote sensing, and carried out experimental verification at the site of the National Forest emission reduction project in Southern China. We used step-up regression and random forest (RF) to screen LiDAR and Landsat 8 OLI multispectral indicators suitable for the prediction of plantation stock, and constructed a plantation stock model based on machine learning (support vector machine regression, RF regression). Our method is compared with traditional statistical methods (stepwise regression and partial least squares regression). Through the verification of 57 plantation field survey data, the accuracy of the stand estimation model constructed using the RF method is generally better (ΔR2 = 0.01~0.27, ΔRMSE = 1.88~13.77 m3·hm−2, ΔMAE = 1.17~13.57 m3·hm−2). The model evaluation accuracy based on machine learning is higher than that of the traditional statistical method, and the fitting R2 is greater than 0.91, while the fitting R2 of the traditional statistical method is 0.85. The best fitting models were all support vector regression models. The combination of UAV point clouds and satellite multi-spectral images has the best modeling effect, followed by LiDAR point clouds and Landsat 8. At present, this method is only applicable to artificial forests; further verification is needed for natural forests. In the future, the density and quality of higher clouds could be increased. The validity and accuracy of the method were further verified. This paper provides a method for predicting the accumulation of typical Chinese plantations at the forest farm scale based on the “airborne LiDAR + satellite remote sensing” data-driven machine learning modeling, which has potential application value for the current carbon neutrality goal of the southern plantation forest emission reduction project.

SEMI-AUTOMATIC ALGORITHM FOR CONSTRUCTING THE CONTOUR OF AREA OBJECTS ON MULTISPECTRAL SATELLITE IMAGE

This article formalizes the problem of semi-automatic construction of the contour of area objects from satellite multispectral images and presents a solution algorithm using PCA and Dijkstra's algorithm. The emphasis of the article is on optimizing PCA using annual matrices for images of the same class, as well as optimizing the Dijkstra algorithm using dynamic programming methods. The contour is considered as the boundary of an object, which can be used for its segmentation and classification. The semi-automatic contour accepts reference points specified by the operator. The formalization of the algorithm is completed.

Dynamic monitoring of surface area and water volume of reservoirs using satellite imagery, computer vision and deep learning

The development of surface area and estimated volume measurements for water reservoirs using advanced computational techniques has become an increasingly important task for reservoir operation. Thus, the objective of this study is to determine the surface area and estimate the volume of water reservoirs through the segmentation of multispectral satellite images using a Fully Convolutional Neural Network (FCN). This paper introduces a universally applicable framework that utilizes deep convolutional neural networks for segmenting water bodies, particularly focusing on areas in Paraíba State, Northeast Brazil, susceptible to severe droughts. The study utilized Sentinel-2 satellite data obtained from Google Earth Engine for the period between January 2019 and September 2023. The extraction of water surface area from satellite images was achieved using the U-Net model, a deep learning framework based on FCN. The accuracy of water surface area extraction was analyzed using the Intersection-Over-Union (IoU) metric, also known as Jaccard Index. Subsequently, reservoir volumes were calculated using elevation-area-volume curves. The evaluation metrics for the results included root mean square error (RMSE), mean absolute percentage error (MAPE), coefficient of determination (R2), and Pearson correlation coefficient (r). The network demonstrated high accuracy, achieving a 95% IoU on the test set. This methodology was applied to four reservoirs, yielding promising results that closely aligned the monitored reservoir curves derived from satellite imagery with observed field data. The findings of this study are significant for research on semiarid region changes, extending their relevance to a broader understanding of global environmental dynamics. This work not only contributes to the field of remote sensing in hydrology but also offers practical insights for effective water resource management in drought-prone regions.

Multispectral Satellite Datasets for Detection and Mapping of Land Cover Change in a Mediterranean Area of North Africa, Belezma (Algeria)

Multispectral satellite images are valuable sources of information for ecologists, agriculture climatologists, or environmentalists. Handling satellite images represents a challenge for better understanding the global change. The Belezma massif by its exceptional geographical position in the Mediterranean region of Northeast Algeria constitutes a fragile hot spot ecosystem with all the degradation factors that directly affect all landcovers. The objective of this study is to determine the contribution of satellite images in the detection of land cover global changes between 1986 and 2021 and landscape monitoring in Belezma region. The methodology adopted is the spatial remote sensing of the evolution of forest stands and the change in land cover, from the processing of Landsat images (L05 TM 1986, L07 ETM+ 2001 and L08 OLI_TIRS 2021) and sampling work for confrontation with the reality on the ground. The outcomes of this investigation demonstrate a general regression rate in the natural vegetation cover that the cedar forest, the steppe, and the wetlands lost 42%, 49%, and 97%, respectively, of their areas between 1986 and 2021, while the degraded and anthropic occupations have increased in favor of market gardening, bare soil, urban areas, and arboriculture by 836%, 158%, 131%, and 50%, respectively, over the same period. This study made it possible to identify and analyze 13 main land uses and plant formations over a period of 35 years, which can be used as a decision-making aid model for the managers of these territories.

METHODS FOR SATELLITE DERIVED BATHYMETRY FROM SENTINEL-2 IMAGES: A COMPARISON

Abstract. In recent decades, research has been developed to estimate near-shore bathymetry depth values using satellite imagery. Visible and infrared bands are used to derive elevation profile estimates, so to obtain bathymetric in rapid way without mobilisation of persons or equipment and saving the costs. For consequence, Satellite Derived Bathymetry (SDB) is seen as a valid approach for shallow waters survey: strongly supported by the activity of scholars and researchers, multiple methods are available in the literature. This article aims to investigate and compare different SDB methods for sea depth extraction from Sentinel-2 satellite multispectral images, with particular attention to the accuracy of the results. The experiments are conducted on imagery including Blue, Green, Red and Near Infrared bands, with 10 m resolution, concerning the Bay of Pozzuoli (Italy). After removing the glint, the effects caused by the reflection of sunlight through single scattering from sea surface, three methods are applied: Band Ratio method (BRM), 3rd-degree polynomial regression line method (3DPM), and principal component analysis method (PCAM). 3DPM can be seen as a variant of the BRM where the linear law that interprets the correlation between the band ratio values and the depth values is replaced by the third order function. Models are trained using depth data extracted from an Electronic Navigational Chart (ENC) at 1:7,500 scale, which is also used to verify result accuracy. The experiments demonstrate that the 3DPM is better able to obtain a more precise bathymetric model, confirming the greater adaptability of the 3rd order function to interpretate the variability of the interaction of light with water along the water column.

A One-Class Classifier for the Detection of GAN Manipulated Multi-Spectral Satellite Images

The current image generative models have achieved a remarkably realistic image quality, offering numerous academic and industrial applications. However, to ensure these models are used for benign purposes, it is essential to develop tools that definitively detect whether an image has been synthetically generated. Consequently, several detectors with excellent performance in computer vision applications have been developed. However, these detectors cannot be directly applied as they areto multi-spectral satellite images, necessitating the training of new models. While two-class classifiers generally achieve high detection accuracies, they struggle to generalize to image domains and generative architectures different from those encountered during training. In this paper, we propose a one-class classifier based on Vector Quantized Variational Autoencoder 2 (VQ-VAE 2) features to overcome the limitations of two-class classifiers. We start by highlighting the generalization problem faced by binary classifiers. This was demonstrated by training and testing an EfficientNet-B4 architecture on multiple multi-spectral datasets. We then illustrate that the VQ-VAE 2-based classifier, which was trained exclusively on pristine images, could detect images from different domains and generated by architectures not encountered during training. Finally, we conducted a head-to-head comparison between the two classifiers on the same generated datasets, emphasizing the superior generalization capabilities of the VQ-VAE 2-based detector, wherewe obtained a probability of detection at a 0.05 false alarm rate of 1 for the blue and red channels when using the VQ-VAE 2-based detector, and 0.72 when we used the EfficientNet-B4 classifier.

Generation of synthetic generative adversarial network-based multispectral satellite images with improved sharpness

Generation of synthetic generative adversarial network-based multispectral satellite images with improved sharpness

Postfire Forest Regrowth Algorithm Using Tasseled-Cap-Retrieved Indices

Wildfires are a common disturbance factor worldwide, especially over the last decade due to global climate change. Monitoring postfire forest regrowth provides fundamental information needed to enhance the management and support of ecosystem recovery after fires. The purpose of this study is to propose an algorithm for postfire forest regrowth monitoring using tasseled-cap-derived indices. A complex approach is used for its implementation, for which a model is developed based on three components—Disturbance Index (DI), Vector of Instantaneous Condition (VIC), and Direction Angle (DA). The final product—postfire regrowth (PFIR)—allows for a quantitative assessment of the intensity of regrowth. The proposed methodology is based on the linear orthogonal transformation of multispectral satellite images—tasseled cap transformation (TCT)—that increases the degree of identification of the three main components that change during a fire—soil, vegetation, and water/moisture—and implies a higher accuracy of the assessments. The results provide a thematic raster representing the intensity of the regrowth classes, which are defined after the PFIR threshold values are determined (HRI—high regrowth intensity; MRI—moderate regrowth intensity; and LRI—low regrowth intensity). The accuracy assessment procedure is conducted using very-high-resolution (VHR) aerial and satellite data from World View (WV) sensors, as well as multispectral Sentinel 2A images. Three different forest test sites affected by fire in Bulgaria are examined. The results show that the classified thematic raster maps are distinguished by a good performance in monitoring the regrowth dynamics, with an average overall accuracy of 62.1% for all three test sites, ranging from 73.9% to 48.4% for the individual forests.

Multi-decadal coastal change detection using remote sensing: the Mediterranean coast of Egypt between El-Dabaa and Ras El-Hekma

A key source of information for many decision support systems is identifying land use and land cover (LULC) based on remote sensing data. Land conservation, sustainable development, and water resource management all benefit from the knowledge obtained from detecting changes in land use and land cover. The present study aims to investigate the multi-decadal coastal change detection for Ras El-Hekma and El-Dabaa area along the Mediterranean coast of Egypt, a multi-sectoral development area. Besides, the superiority of the area is highly dependent on its proximity to three development projects: the tourism and urban growth pole at Ras El-Hekma, the beachfront Alamain New Mega City, and the Nuclear Power Plant at El Dabaa. This study utilized multi-spectral Landsat satellite images covering 1990, 2010, and 2020 to perceive the post-classification change detection analysis of the land use and land cover changes (LULCC) over 30 years. The results of the supervised classification from 1990 to 2020 showed a 47.33 km2 (4.13%) expansion of the agricultural land area, whereas the bare soil land area shrunk to 73.13 km2 (6.24%). On the other hand, the built-up activities in the area launched in 2010 and escalated to 20.51 km2(1.77%) in 2020. The change in land use reveals the shift in the economic growth pattern in the last decade toward tourism and urban development. Meanwhile, it indicates that no conflict has yet arisen regarding the land use between the expanded socioeconomic main sectors (i.e., agriculture, and tourism). Therefore, the best practices of land use management and active participation of the stakeholders and the local community should be enhanced to achieve sustainability and avoid future conflicts. An area-specific plan including resource conservation measures and the provision of livelihood alternatives should be formulated within the National Integrated Coastal Zone Management (ICZM) plan with the participation of the main stakeholders and beneficiaries. The findings of the present work may be considered useful for sustainable management and supportive to the decision-making process for the sustainable development of this area.

Urban hydrological model (UHM) developed for an urban flash flood simulation and analysis of the flood intensity sensitivity to urbanization

An urban hydrological model (UHM) forced by radar-observed rainfalls was developed for producing rainstorm-related urban inundation maps for obtaining flash flood forecasts. The parameter sets of the land use and land cover (LULC) and urban drainage capacity derive from satellite multispectral images and high spatial resolution GIS datasets, relating to the urban hydrology and hydraulic properties. The hydrodynamics model was based on the simplified shallow water equation (SWE) neglecting the convective acceleration and pressure terms in the St. Venant equation. The intense convective rainstorm that induced heavy flash floods on 21 Jul. 2012 and the large-scale stratiform precipitation that occurred from 19 Jul. to 21 Jul. 2016 in Beijing were deliberately selected to perform flood-simulation and model-validation case studies. The simulation of severe flooding scenarios in the 2012 event was fairly reproduced and verified using media reports. In the case studies, the sensitivity tests verified that the flood intensity (1) would be enhanced slightly on pure impervious surfaces, (2) would be reduced significantly on pure pervious surfaces, and (3) would increase by 30%–60% without a drainage system. No false alarm could be perceived in the simulation of the 2016 event in a prolonging stratiform precipitation. The case studies confirmed the perspectives of using the hydrological model fed by QPEs/QPFs for flash flood forecasting and concluded that the hourly precipitation surpassing 30 mm/h would be indicative in impending flash floods in Beijing metropolis.

The Extended Stumpf Model for Water Depth Retrieval From Satellite Multispectral Images

The Extended Stumpf Model for Water Depth Retrieval From Satellite Multispectral Images

Analysis of land use change using remote sensing and GIS techniques in the Cazaderos forest, Loja province, Southern Ecuador

The predominant aim of the current study has been to determine land use change over a period of eight years in the Cazaderos dry forest in the province of Loja in Ecuador using remote sensing and GIS techniques. The methodology employed included the application of supervised classification of multispectral satellite imag-es applied to freely available Landsat 8 images using Catalyst image processing software and the application of the Object Analysis technique, which is presented as the methodological protocol in remote sensing that offers the highest level of accuracy. This process was accompanied by complementary Geographical Infor-mation System (GIS) techniques that allowed spatialization and quantification of the resulting information. The Cazaderos dry forest in the period of analysis has suffered a significant spatial fragmentation dominated by the current presence of bare soils that were previously covered by dense shrubs and have undergone a process of deforestation. During the period 2013 - 2021 the deforestation process extended to an area of 1273.95 ha corresponding to 15.21% of the total surface of the study area while reforestation reached 1426.86 ha corresponding to 17.04% of the total surface.

An assessment of seasonal changes in forests using the segmentation and cluster analysis of Landsat space images

For the purposes of monitoring the state of forest ecosystems, it is most effective to use the capabilities of remote methods. On multispectral Landsat satellite images (a time series of summer and autumn images for 2000, 2001, 2013, 2014) of the territory of the Olekminsky State Nature Reserve, fragments of forests dominated by Gmelin larch (Larix gmelinii Rupr.) with an area of 250 km2 (scale 1:5000). Then the polygons were saved at three levels of segmentation (scaling) – 4, 16, 64 with scales of 1:2500, 1:1250, 1:625. During decryption, uncontrolled classification of polygons using the ISODATA (Iterative Self-Organizing Data Analysis Technigue) method into 2,4,10 classes was carried out. Classification into two classes was used to calculate the forest cover index. Seasonal changes were determined by the difference in the forest cover index values of the polygons in summer and autumn. It has been shown that the greater the difference, the greater the proportion of larch in mixed forest stands. Distribution curves of forest cover index values were constructed for polygons of the 3rd scaling level. Based on the results of classification into 4, 10 classes, statistical processing was carried out with the calculation of indicators of difference and similarity of polygons - dispersion of the general population and Fisher's test (F-test). The results of changes in population dispersion and F-test at different levels of segmentation and in different years are considered.

Multiscale Fusion of Panchromatic and Multispectral Images Based on Adaptive Iterative Filtering

This paper proposes an efficient and high-fidelity image fusion method based on adaptive smoothing filtering for panchromatic (PAN) and multispectral (MS) image fusion. The scale ratio reflects the ratio of spatial resolution between the panchromatic image and the multispectral image. When facing a multiscale fusion task, traditional methods are unable to simultaneously handle the problems of spectral resolution loss resulting from high scale ratios and the issue of reduced spatial resolution due to low scale ratios. To adapt to the fusion of panchromatic and multispectral satellite images of different scales, this paper improves the problem of the insufficient filtering of high-frequency information of remote sensing images of different scales by the classic filter-based intensity modulation (SFIM) model. It uses Gaussian convolution kernels instead of traditional mean convolution kernels and builds a Gaussian pyramid to adaptively construct convolution kernels of different scales to filter out high-frequency information of high-resolution images. It can adaptively process panchromatic multispectral images of different scales, iteratively filter the spatial information in panchromatic images, and ensure that the scale transformation is consistent with the definition of multispectral images. Using 15 common fusion methods, this paper compares the experimental results of ZY-3 with scale ratio 2.7 and SV-1 with scale ratio 4 data. The results show that the method proposed in this paper retains good spatial information for image fusion at different scales and has good spectral preservation.