Analyze Changes Using CCDC (Image Analyst)

Disponible avec une licence Image Analyst.

Résumé

Évalue les modifications dans les valeurs de pixel au fil du temps à l’aide de la méthode CCDC (Continuous Change Detection and Classification) et génère un raster multidimensionnel contenant les résultats du modèle.

Learn more about how Analyze Changes Using CCDC works

Utilisation

• The Continuous Change Detection and Classification (CCDC) algorithm is a method for identifying change in pixel values over time. It was originally developed for a time series of multiband Landsat imagery and is used to detect change and classify land cover before and after the change occurred. This tool can be used with imagery from supported sensors and can also be used to detect change in single band rasters. For example, this tool can be used to detect changes in a time series of NDVI rasters to identify deforestation events.

• The input multidimensional raster must have at least 12 slices, spanning at least 1 year.

• It is recommended that you remove cloud and cloud shadow mask using a QA band before running this tool.

• To explore the changes calculated in the output change analysis raster, create a temporal profile chart. Generate graphs for various locations in the change analysis raster to see when changes have occurred. For pixels that have changed, the graph will show breaks in the model where the harmonic regression model for the pixel values over time shifted to a new model, indicating a change. You can hover the pointer over the points on the graph to identify the date the model changed.

• The output change analysis raster is a multidimensional raster in which each slice is a multiband raster composed of the time series model coefficients, root mean square error (RMSE), and the observed changes. The number of slices in the output matches the number of slices in the input. It can be used as the input to the Detect Change Using Change Analysis Raster tool, which generates a raster containing change information for each pixel.

• The output change analysis raster can also be used for classification. Run this tool to generate a change analysis raster. Then create training samples with a time field to indicate the time at which the sample represents land cover. Next, run a training tool to generate a classifier definition file (.ecd). Finally, run the Classify Raster tool with the .ecd file and the change analysis raster as inputs to generate a multidimensional classified raster.

• Le paramètre Bands for Temporal Masking (Bandes pour le masquage temporel) spécifie les canaux à utiliser pour le nuage, l’ombre des nuages et le masquage de la neige. L’ombre des nuages et la neige apparaissant très sombres dans le canal à onde courte infrarouge (SWIR) et les nuages et la neige étant très lumineux dans le canal vert, il est recommandé de masquer les index de canaux des canaux SWIR et vert.

• Le paramètre Updating Fitting Frequency (in years) (Mise à jour de la fréquence de pertinence (en années)) définit la fréquence à laquelle le modèle de série chronologique est mis à jour avec les nouvelles observations. La mise à jour fréquente d’un modèle peut consommer une grande quantité de ressources informatiques pour un avantage minimal. Par exemple, si le raster multidimensionnel contient 365 tranches ou observations nettes par an et que la fréquence de mise à jour concerne chacune de ces observations, le traitement demandera 365 fois plus de ressources informatiques qu’une mise à jour annuelle, sans pour autant que la précision ne soit supérieure.

• This tool may take a long time to run and requires significant disk space to store the results. To improve processing time and reduce the amount of storage space, the following steps are recommended:

  • Turn off the Pyramid environment. Uncheck the Build Pyramids box in the Environment pane or set the environment to NONE in Python.
  • Set the Compression environment to LERC and set Max error to 0.000001.
  • If you expect to run the Detect Change Using Change Analysis Raster tool on the output of this tool multiple times, it is recommended that you build a multidimensional transpose on the result.

• Cet outil crée un jeu de données raster multidimensionnelles au format CRF (Cloud Raster Format). Actuellement, aucun autre format en sortie n’est pris en charge.

Syntaxe

AnalyzeChangesUsingCCDC(in_multidimensional_raster, {bands}, {tmask_bands}, {chi_squared_threshold}, {min_anomaly_observations}, {update_frequency})

Paramètre	Explication	Type de données
in_multidimensional_raster	Jeu de données raster multidimensionnelles en entrée.	Raster Dataset; Raster Layer; Mosaic Dataset; Mosaic Layer; Image Service
bands [bands,...] (Facultatif)	The band IDs to use for change detection. If no band IDs are provided, all the bands from the input raster dataset will be used.	Long
tmask_bands [tmask_bands,...] (Facultatif)	The band IDs to be used in the temporal mask (Tmask). It is recommended that you use the green band and the SWIR band. If no band IDs are provided, no masking will occur.	Long
chi_squared_threshold (Facultatif)	The chi-square statistic change probability threshold. If an observation has a calculated change probability that is above this threshold, it is flagged as an anomaly, which is a potential change event. The default value is 0.99.	Double
min_anomaly_observations (Facultatif)	The minimum number of consecutive anomaly observations that must occur before an event is considered a change. A pixel must be flagged as an anomaly for the specified number of consecutive time slices before it is considered a true change. The default value is 6.	Long
update_frequency (Facultatif)	The frequency, in years, at which to update the time series model with new observations. The default value is 1.	Double

Valeur renvoyée

Nom	Explication	Type de données
out_ccdc_result	Jeu de données raster multidimensionnelles CRF (Cloud Raster Format) en sortie. The output change analysis raster containing model information from the CCDC analysis.	Raster

Exemple de code

AnalyzeChangesUsingCCDC example 1 (Python window)

This example performs continuous change detection on 30 years of monthly NDVI rasters. Only one band is used in the change detection and the chi-squared probability threshold is 0.90.

# Import system modules
import arcpy
from arcpy.ia import *

# Check out the ArcGIS Image Analyst extension license
arcpy.CheckOutExtension("ImageAnalyst")


changeAnalysisRaster = arcpy.ia.AnalyzeChangesUsingCCDC(
	"Monthly_NDVI_30_years.crf", [0], [], 0.90, 6, 1); 

# Save output
changeAnalysisRaster.save(r"C:\data\NDVI_ChangeAnalysis.crf")

AnalyzeChangesUsingCCDC example 2 (stand-alone script)

This example performs continuous change detection on a time series of Landsat 7 images, with bands 3 and 7 (indexed at 2 and 6) used for a snow, cloud, and cloud shadow mask.

# Import system modules
import arcpy
from arcpy.ia import *

# Check out the ArcGIS Image Analyst extension license
arcpy.CheckOutExtension("ImageAnalyst")

# Define input parameters
in_multidimensional = r"C:\data\Landsat_time_series.crf"
change_bands = [0,1,2,3,4,5,6]
tmask_bands = [2,6]
chi_sq_threshold = 0.99
min_consecutive_observations = 3
update_frequency = 1

# Execute
changeAnalysisRaster = arcpy.ia.AnalyzeChangesUsingCCDC(
	in_multidimensional, change_bands, tmask_bands, chi_sq_threshold, 
	min_consecutive_observations, update_frequency) 

# Save output
changeAnalysisRaster.save(r"C:\data\Landsat_ChangeAnalysis.crf")

Environnements

Taille de cellule, Compression, Espace de travail courant, Transformations géographiques, NoData, Mot-clé de configuration en sortie, Système de coordonnées en sortie, Etendue, Facteur de traitement parallèle, Pyramide, Statistiques raster, Méthode de ré-échantillonnage, Espace de travail temporaire, Raster de capture, Taille de tuile

Informations de licence

• Basic: Requiert Image Analyst
• Standard: Requiert Image Analyst
• Advanced: Requiert Image Analyst