Lab 3: Unsupervised Classification

Goal
The main goal of this lab is to understand the art of extracting biophysical and sociocultural information from remotely sensed images, utilizing unsupervised classification algorithm. First in the lab we will become familiar with understanding the input configuration requirements and execution of an unsupervised classifier. Then the art of recoding multiple spectral clusters generated from unsupervised classifiers into thematic informational land use and land cover classes that meet a classification scheme.
Methods
Part 1: Experimenting with unsupervised ISODATA classification algorithm  
Section 1: Setting up an unsupervised classification algorithm  

 In part one of the lab, we will  be experimenting with unsupervised ISODATA classification algorithm. First off, an Eau claire and Chippewa Image was formated to an unsupervised classification, utilizing an ISODATA algorithm.

Figure 1: Classification scheme utilized when performing the ISODATA unsupervised classification.

 Section 2: Labeling of unsupervised clusters into meaningful land use/land cover classes 
Once this image was produced, then the image was synced to a google Earth viewer so the clusters of data could be categorized into the 5 classifications showed in figure 1.When these classifications were applied after determining what the surface feature was, the attribute is as follows in Figure 2.


Figure 2: This attribute table is what i classified each cluster to be. based on analyzing different values.

Part 2: Improving the accuracy of unsupervised classification
Section 1: Setting up and running an unsupervised classification algorithm  
In the second part of this lab, the same process was performed as in part 1 section 1 with the Eau Chipp2000 image. However the maximum number of classes this time is 20 rather than 10. The convergence threshold this time was lowered to 0.92 from 0.95. This threshold is the lowest number of pixel values that are considered within each classification.  The output image (Eau_Chipp2000usp2.img) now has 20 clusters of data in the attribute table (Figure 3). This data will be classified and renamed to fall under one of the 5 categories mentioned above in part 1. Figure 4 shows the classification of the 20 new classes.

Figure 3: This attribute table shows the new ISODATA classification with 20 clusters, and 0.92 threshold.

Results
Part 1: Experimenting with unsupervised ISODATA classification algorithm  
In part one, Figure 4 shows the result of performing unsupervised ISODATA classification. This classification is not very accurate. The green color represents forest, however there's not that much forest in the Eau Claire and Chippewa Counties. additionally faulty classifications are seen in all classes, agriculture especially because the Eau Claire Airport Runway is considered to be Agricultural land and not urban. This is problematic because classification maps should be more accurate.
Figure 4: This image shows the ISODATA classification scheme with 20 classes symbolized as 5.
Part 2: Improving the accuracy of unsupervised classification
In part 2 when the LULC classification grouped the 20 cluster into 5 classes, the final output is shown in Figure 5. This improvement of the unsupervised classification is slightly more accurate i believe, except there is to much user input to know if the user was correct in naming each class. 
Figure 5: This map shows 20 classes clustered into 5 to show a slightly more accurate classification.
Sources

Landsat satellite image is from Earth Resources Observation and Science Center, United States Geological Survey. 

Comments

Post a Comment

Popular posts from this blog

Lab 6: Digital Change Detection

Image
Background The goal of this lab is to develop an image that shows change detection of land surface features over time. First a means of performing a qualitative change detection will be performed, then we will quantify post-classification change detection, and a model will be developed that will map detailed change of land use land cover images over time. Methods Part 1: Change detection using Write Function Memory Insertion The first part of the lab, a Write Function Memory Insertion is utilized. This processes utilizes near infrared bands, when the two images are brought into the Write Function, an output image is produced, and the image has areas highlighted in pinkish red, these are areas that have experienced change. Specifically in this lab, ec_envs1991.img and ec_envs_2011.img were used to perform this change detection. Part 2: Post-classification comparison change detection   In this part of the lab a From-to change detection using classified images of the Milw...
Read more

Lab 5: classification Accuracy Assessment

Image
Goal The goal of this lab is to understand how to evaluate the accuracy of classification results. Accuracy assessments are mandatory after performing an image classification, this is a portion of the post-processing stage of remotely sensed data. Methods An accuracy assessment is performed on a supervised and an unsupervised classification from labs 3 and 4. This is done by utilizing the accuracy assessment tool in Erdas. A high resolution reference image of the same study area is used to perform the assessment. This is done by generating  random points on the reference image, 125 random points in this lab,  and classifying those points based on the high resolution image utilizing the same classification codes for water, forest, agriculture, etc. (Table 1). By identifying these points, a comparison report is then produced (Figure 1). This allows for an error matrix to be created from the results of comparing the random points to the classified images. Table 1: ...
Read more

Lab 10: Radar Remote Sensing

Image
Goal The goal of this lab is to understand how to perform basic preprocessing and processing of radar images. Specifically noise reduction through speckle filtering, spectral and spatial enhancement, multi-sensor fusion, texture analysis, polarimetric processing, and slant-range to ground-range conversion. When this lab is completed, the basic knowledge of processing on any type of synthetic aperture radar image will be gained. Methods Speckle Supression In this lab radar imagery is analyzed and processed through Erdas Imagine, and ENVI. Noise reduction methods for preprocessing of radar imagery was performed through speckle filtering, edge enhancement and image enhancement. To perform speckle filtering, Radar speckle suppression tool is used. this removes common salt and pepper effect. this is done by utilizing a calculated coefficient of variation (Figure 1). Figure 1: coefficient of variation is 0.274552. This variable is then used to run the Speckle Suppression Functio...
Read more