The spatiotemporal remote sensing NDVI dataset and spatiotemporal analysis were constructed in ENVI

Mondo Science Updated on 2024-03-06

We often use multi-time satellite remote sensing imagery to analyze some characteristic changes in an area, such as urban expansion, vegetation changes, temperature changes, dynamic cloud maps, typhoon evolution**, etc. In this paper, we introduce the construction of spatiotemporal remote sensing datasets in ENVI, and use the band tool and pixel profile tool for spatiotemporal analysis. The following is an animation of the evolution of NDVI generated from multitemporal and multispectral images.

envi has two tools that enable you to build spatiotemporal files that can be used for analysis:

Build Temporal Cube: Build a multi-band raster file, where each band represents a different date-time (temporal cube).

Build Raster Series: Build a series of images into a JSON file and view the images incrementally (time raster).

The two tools look basically the same, but there are actually big differences. The following is an example of multi-view Landsat data to describe the use of several tools. Remote sensing images applied to time series analysis generally need to include the acquisition time, which is generally available for data opened and processed in ENVI, and can be manually added by the following methods if not:

1) Open the data to which you want to add the time attribute.

2) Open View Metadata and click Edit Metadata to open the metadata editor.

3) Click the Add button in the upper left corner to add the acquisition time attribute and select the corresponding time information.

Build temporal cube tool

If the input is a multispectral image, you need to select a spectral index to calculate a spectral index for each input multispectral image, and generate the calculated spectral index to generate a space-time cube file output, and if the input is a single-band image, the space-time cube file output is directly generated. A space-time cube file is essentially a multi-band raster file.

1) Open multi-phase multispectral data.

2) In the toolbox, open the SpatioTemporal Analysis Build Temporal Cube. In the file selection dialog box, select multi-temporal and multispectral data, and hold shift or ctrl to select multi-spectral data.

3) In the Build Temporal Cube panel that opens, set the following parameters.

Overlap Operation: The method of handling overlap areas, including intersection and union.

spectral index: Select a spectral index. When the input data is a multispectral image, it must be selected, and the tool automatically lists the spectral indices supported by the band.

date format: The name of the generated band is in the time command, here select the time format.

Figure: Build Temporal Cube panel.

4) Select the output file name and path, or choose to export to virtual raster.

Build Raster Series tool

The Build Raster Series tool is to record the input multi-phase image file and generate one. A text file in JSON format with a series suffix. There are no excessive requirements for the input multi-phase image file, which can have coordinate information or pixel coordinates.

1) There are two ways to launch the tool:

From Toolbox, select Spatiotemporal Analysis > Build Raster Series.

From the menu, select File > New > Build Raster Series.

2) Open the Build Raster Series panel, there are several ways to add data:

add by filename: Select a multiphase file from the local path. Use the Ctrl or Shift keys to multi-select.

add open raster: Select a file from the list of open data.

add by filterSelect files in bulk from folders.

Note: Others can be added. series file, click itimport seriesButton.

3) If there is a time attribute in the input multi-phase image, check Order by Time will automatically sort by time, otherwise it will be sorted by the input file.

4) Select the output file name and path, and click OK to generate the timing file.

Figure: Build Raster Series panel.

The time series file obtained by the Build Raster Series tool does not make any modifications to the original multi-temporal image, that is, the multi-temporal image that makes up the time series file can be different coordinate information, different ranges, etc., for example, the same spatial range is not required to track typhoons using multi-temporal animation. When you need to normalize multi-temporal images to the same range or coordinates, the following tools can post-process the generated time series files:

Regrid Raster Series: You can customize the output time series file in a variety of ways, including coordinate type, spatial extent, cell size, and more.

regrid raster series by index: defines the output timing file by one of the timing files.

regrid raster series by intersection: Defines the output timing file by the intersection of the individual files in the timing file.

regrid raster series by union: defines the output timing file by the union of the individual files in the timing file.

Spatiotemporal analysis

Spatiotemporal analysis is mainly done using the Band Animation tool or the Series Animation Manager, which have similar functions and can be opened in the following ways:

In the Layer Manager, right-click a layer and select Band Animation or Display > Band Animation.

display > series animation manager or file > open as > series.

Figure: Band Animation tool.

The following describes the use of the build temporal cube tool as an example.

1) The Build Temporal Cube tool generates a multi-phase NDVI spatiotemporal dataset and displays the file.

2) After opening the band animation tool, a temporary layer will be generated in the layer manager, right-click on this layer to change the color table, and select a color for rendering.

3) Use the annotation in the toolbar to add some annotations, such as text, kilometer net, scale, color band and other information.

4) On the band animation tool, click the button and select annotate->band, you can add the band name as a dynamic annotation, where the band name is the imaging time.

5) Through the Band Animation tool, you can animate each band.

6) Click the button, select S**e Video Animation, and save the animation as a file in **i, gif, mp4 and other formats.

A similar approach can be used to analyze the timing file obtained by the Build Raster Series tool to obtain a true color animation.

Figure: True color ** animation.

The NDVI Data is shared as follows:

(1) Personnel, limited to users of interest.

(2) All kinds of projects (including all kinds of scientific research projects) apply for this data throw to enjoy the free policy, but you need to donate a certain number of hard disks to this number to obtain it.

(3) Donate a hard disk to get data without leaving a message.

Like, share, followWe get the data!!

What else to see? Hurry up and make a tripleandClick on the original articleGet the data!

Related Pages