Evaluation of algorithms for the fusion of multispectral and panchrome images in python

 



In the conducted research, three fusion algorithms were evaluated for combining the panchromatic image with the RGB bands of Landsat 8 satellite images. The analyzed algorithms were Brovey, EIHS, and Averaging. The objective was to determine which algorithm produces the best results in terms of spatial resolution enhancement and visual quality of the final image.

The Brovey algorithm is based on the intensity relationship between the panchromatic and RGB bands. This method assigns a weight to each RGB band based on its relative contribution to the total intensity of the original image. The panchromatic band is then fused with the weighted RGB bands to obtain a final image.

On the other hand, the Enhanced Intensity Hue Saturation (EIHS) algorithm focuses on improving the visual quality of the final image. This method preserves the spectral information of the RGB bands while enhancing the spatial resolution using the panchromatic band. The fusion is performed by transforming the original image to the HSV color space, where a modification is made to the intensity component using the information from the panchromatic band.

Lastly, the Averaging algorithm employs a simpler and more direct approach. It involves averaging the pixel values of the panchromatic band with the corresponding pixel values of the RGB bands. This method does not consider the intensity relationship between the bands and mainly focuses on improving spatial resolution.

To evaluate the algorithms, various evaluation metrics such as the ERGAS and RMSE were utilized. Visual tests were also conducted to assess the quality of the fused image in terms of sharpness, contrast, and color fidelity.

The evaluation results indicated that the Brovey algorithm exhibited a significant improvement in spatial resolution of the fused image but sometimes introduced visual artifacts and spectral information loss. On the other hand, the EIHS algorithm managed to enhance the visual quality of the final image while preserving spectral information but showed a lesser improvement in spatial resolution compared to the Brovey algorithm. The Averaging algorithm yielded acceptable results in terms of spatial resolution and visual quality, albeit without standing out in either category.

Comments

Post a Comment

Popular posts from this blog

MAIZAPP

Image
  Common rust is a fungal disease that affects maize crops and can have a significant impact on agricultural production. Early detection of this disease is crucial for taking preventive measures and mitigating its spread. In this context, a mobile application has been developed that utilizes computer vision techniques and convolutional neural networks, particularly the YOLO (You Only Look Once) model, to identify the presence of common rust on maize leaves. The process of creating the application consists of several stages. Firstly, images of maize leaves affected by common rust are collected and labeled, along with images of healthy leaves as reference. These images are used to train the YOLO model, which is a specialized neural network for object detection and localization in images. During training, the model learns to recognize the visual patterns associated with common rust on maize leaves. Once the training is complete, the model is optimized and adjusted for implementation o...
Read more

Landsat 8 Satellite Image Editor

Image
  This text describes the creation of a graphical user interface (GUI) in Python for editing satellite images. The interface will include several functions, including enhancing resolution, removing cloud cover, cropping a region of interest, and recording band information. Enhancing resolution: This function allows increasing the resolution of a satellite image to obtain more details. Techniques such as interpolation or advanced algorithms for resizing the image can be applied Removing cloud cover: Clouds in satellite images can hinder interpretation and analysis. This function utilizes image processing algorithms to detect and remove clouds, improving the visibility of the underlying image. Cropping a region of interest : This function allows the user to select a specific region of the satellite image for cropping and saving it as an independent image. Users can specify coordinates or use an interactive tool to define the region of interest. Recording band information: Satellit...
Read more

Geographic Viewer Cadastre

Image
  This text presents an interactive geovisualizer designed to manage the physical appraisal of properties and perform economic analysis in a specific locality. The tool utilizes a combination of programming languages, including HTML, JavaScript, CSS, Overleaf, and PostGIS, to provide an intuitive interface and advanced functionalities. The geovisualizer allows users to visualize geographic maps of the locality, overlaying layers of relevant information such as cadastral data, infrastructure, public services, development zones, among others. These layers are generated and updated using PostGIS, an extension of the PostgreSQL database specialized in handling spatial data. Through the geovisualizer interface, users can select a specific property and access its detailed information, including physical characteristics, dimensions, current use, and historical appraisal data. Additionally, a functionality is provided to automatically perform a physical appraisal of the property, using alg...
Read more