H25+V2

Tile H25 V2 covers far Eastern Russia and has areas that have experienced both an increase and a decrease in the vegetation. The largest areas of change have occured in the form of an increase in vegetation. These increases could be due to many different factors such as conservation of Siberian Tiger habitat, better preservation of land resources, or even climate change.



Eastern Russia Spring NDVI

Eastern Russia Summer NDVI

Eastern Russia Fall NDVI


 * Lab 4**

Questions: 1. Write down the image dimensions (e.g. number of lines, columns and bands) for the Anthromes file and the MODIS time series. Modis: 121 lines, 144 bands, 1 column Anthromes: 121 lines, 1 band, 296 columns 2. How do the number of lines and columns compare for the time series and the Anthromes file? Are they the same or different, why? How many time steps does your time series have? How many observations per year? 3. Discuss why it would not be right to shrink the Anthrome pixels. Try it out if you want. Shrinking the Anthrome pixels would make it hard to distinguish all the different classes, plus larger is easier to interpret. 4. How much area is covered in your region by the 5 most dominant classes? 94.28% 5. Discuss the differences between NDVI and the coefficient of variation for the 5 classes. 6. Can you determine when the growing season starts and ends for each class? Does it change by year? 7. How do you think the phenology in this biome would have looked if there were no humans? I believe the phenology would look the same because there is very little presence of humans in this area. 8. Do you find anything remarkable in the time series? Are there trends? 9. Are there other things you want to discuss in these graphs?



List the 5 dominant classes in your tile (in order from most dominant to least dominant). 1. 62 Sparse Trees 2. 52 Remote Forests 3. 61 Wild Forests 4. 43 Remote Rangeland 5. 51 Populated Forests
 * # || Class || # pixels in area || % pixels in area ||
 * 11 || Urban || 1 || 0.001 ||
 * 12 || Dense settlements || 1 || 0.001 ||
 * 21 || Rice villages || 0 || 0 ||
 * 22 || Irrigated villages || 0 || 0 ||
 * 23 || Cropped pastoral villages || 0 || 0 ||
 * 24 || Pastoral villages || 3 || 0.004 ||
 * 25 || Rainfed villages || 0 || 0 ||
 * 26 || Rainfed mosaic villages || 2 || 0.003 ||
 * 31 || Residential irrigated cropland || 1 || 0.001 ||
 * 32 || Residential rainfed mosaic || 67 || 0.087 ||
 * 33 || Populated irrigated cropland || 1 || 0.001 ||
 * 34 || Populated rainfed cropland || 33 || 0.043 ||
 * 35 || Remote croplands || 0 || 0 ||
 * 41 || Residential rangelands || 7 || 0.009 ||
 * 42 || Populated rangelands || 37 || 0.048 ||
 * 43 || Remote rangelands || 4613 || 5.994 ||
 * 51 || Populated forests || 2845 || 3.697 ||
 * 52 || Remote forests || 23454 || 30.477 ||
 * 61 || Wild forests || 6868 || 8.925 ||
 * 62 || Sparse trees || 34798 || 45.218 ||
 * 63 || Barren || 2621 || 3.406 ||


 * Lab 5**

Below is the spectral profile for a time-series from Feb. 2, 2001-2008 Below is the spectral profile for a time-series from May 17, 2001-2008 Below is a map of the mean snow for the Feb. time-series Below is a map of the standard deviation for snow in the Feb. time-series

If you look at the spectral profiles and maps you can see that the amount of snow fall over the eight years varies from year to year. Due to the fact that the amount of snow varies so much through the years there is no conclusive evidence to say that the amount of snow has had any affect on the increase in vegetation over the last eight years.

=Landsat=

Lab7

Available Landsat Images with 20% cloud cover or less
 * Date || Cloud Cover || Download ||  ||   ||
 * 7/24/2009 || 0% || No ||  ||   ||
 * 9/29/2010 || 14% || No ||  ||   ||

Lab 9

Gaussian Filter with 25x25 Kernel and 40% Add Back

High Pass Filter with 25x25 Kernel and 40% Add back

Median Filter 25x25 kernel 40% Add Back

Of all three filters that were applied I feel like the Median Filter is the best, however I feel that the true color original looks even better than the Median Filter. The Median Filter appears to be more zoomed in than the rest but also appears to be the most clear than the other filters.Of the two High Pass Filters the Gaussian is the better of the two.The Gaussian Filter is more crisp and shows the vegetation better than the original High Pass Filter.

True Color Original

PC Parameters in order from bands 1-6





Eigenvalues



Percent of variability Band 1 85% Band 2 13% Band 3 0.7%

True Color Image of PC's

Lab 10

Two Color Multiview for band 4



PC depicting change



NDVI Change




 * Lab 11& 12**

Separation Table For Image One Pair Separation (least to most); Cloud Shadow [Black] 1116 points and Water [Blue] 1205 points - 1.46174150 Forest [Green] 1181 points and Plains [Yellow] 1248 points - 1.98895356 Bedrock [Maroon] 1470 points and Cloud Shadow [Black] 1116 points - 1.99504113 Bedrock [Maroon] 1470 points and Cloud [White] 1027 points - 1.99947750 Bedrock [Maroon] 1470 points and Water [Blue] 1205 points - 1.99968633 Bedrock [Maroon] 1470 points and Plains [Yellow] 1248 points - 1.99977142 Bedrock [Maroon] 1470 points and Forest [Green] 1181 points - 1.99994236 Cloud [White] 1027 points and Cloud Shadow [Black] 1116 points - 1.99996514 Forest [Green] 1181 points and Cloud Shadow [Black] 1116 points - 1.99997612 Forest [Green] 1181 points and Cloud [White] 1027 points - 1.99998326 Cloud [White] 1027 points and Plains [Yellow] 1248 points - 1.99998384 Cloud [White] 1027 points and Water [Blue] 1205 points - 1.99999508 Forest [Green] 1181 points and Water [Blue] 1205 points - 1.99999997 Cloud Shadow [Black] 1116 points and Plains [Yellow] 1248 points - 2.00000000 Plains [Yellow] 1248 points and Water [Blue] 1205 points - 2.00000000

Separation Table For Image Two Pair Separation (least to most); Forest [Green] 1996 points and Plains [Yellow] 1579 points - 1.79038434 Rock [Maroon] 1650 points and Cloud Shadow [Black] 1605 points - 1.82440366 Water [Blue] 1385 points and Cloud Shadow [Black] 1605 points - 1.93779363 Rock [Maroon] 1650 points and Water [Blue] 1385 points - 1.97560570 Rock [Maroon] 1650 points and Plains [Yellow] 1579 points - 1.98483641 Rock [Maroon] 1650 points and Forest [Green] 1996 points - 1.99380854 Forest [Green] 1996 points and Cloud Shadow [Black] 1605 points - 1.99630574 Rock [Maroon] 1650 points and Clouds [White] 1807 points - 1.99750213 Plains [Yellow] 1579 points and Clouds [White] 1807 points - 1.99985254 Plains [Yellow] 1579 points and Cloud Shadow [Black] 1605 points - 1.99998495 Forest [Green] 1996 points and Clouds [White] 1807 points - 1.99999856 Water [Blue] 1385 points and Forest [Green] 1996 points - 1.99999995 Clouds [White] 1807 points and Cloud Shadow [Black] 1605 points - 2.00000000 Water [Blue] 1385 points and Plains [Yellow] 1579 points - 2.00000000 Water [Blue] 1385 points and Clouds [White] 1807 points - 2.00000000

(Maroon: Rock; Blue: Water; Green: Forest; Yellow: Plains; White: Clouds; Black: Cloud Shadow/other Shadow)
 * Image One Classification Image Two Classification**

Class Confusion Matrix Image One Confusion Matrix: [Memory1] (1500x1500x1)

Overall Accuracy = (5189/5257) 98.7065% Kappa Coefficient = 0.9844

Ground Truth (Pixels) Class Rock Forest Clouds Cloud Shadow Water Unclassified 0 0 0 0 0 Rock [Maroon] 963 0 0 0 3 Forest [Green 0 787 0 0 10 Clouds [White 0 0 1077 0 0 Cloud Shadow 0 0 0 930 10 Water [Blue] 0 0 0 42 705 Plains [Yello 0 0 0 0 0 Total 963 787 1077 972 728

Ground Truth (Pixels) Class Plains Total Unclassified 0 0 Rock [Maroon] 2 968 Forest [Green 1 798 Clouds [White 0 1077 Cloud Shadow 0 940 Water [Blue] 0 747 Plains [Yello 727 727 Total 730 5257

Ground Truth (Percent) Class Rock Forest Clouds Cloud Shadow Water Unclassified 0.00 0.00 0.00 0.00 0.00 Rock [Maroon] 100.00 0.00 0.00 0.00 0.41 Forest [Green 0.00 100.00 0.00 0.00 1.37 Clouds [White 0.00 0.00 100.00 0.00 0.00 Cloud Shadow 0.00 0.00 0.00 95.68 1.37 Water [Blue] 0.00 0.00 0.00 4.32 96.84 Plains [Yello 0.00 0.00 0.00 0.00 0.00 Total 100.00 100.00 100.00 100.00 100.00

Ground Truth (Percent) Class Plains Total Unclassified 0.00 0.00 Rock [Maroon] 0.27 18.41 Forest [Green 0.14 15.18 Clouds [White 0.00 20.49 Cloud Shadow 0.00 17.88 Water [Blue] 0.00 14.21 Plains [Yello 99.59 13.83 Total 100.00 100.00

Class Commission Omission Commission Omission (Percent) (Percent) (Pixels) (Pixels) Rock [Maroon] 0.52 0.00 5/968 0/963 Forest [Green 1.38 0.00 11/798 0/787 Clouds [White 0.00 0.00 0/1077 0/1077 Cloud Shadow 1.06 4.32 10/940 42/972 Water [Blue] 5.62 3.16 42/747 23/728 Plains [Yello 0.00 0.41 0/727 3/730

Class Prod. Acc. User Acc. Prod. Acc. User Acc. (Percent) (Percent) (Pixels) (Pixels) Rock [Maroon] 100.00 99.48 963/963 963/968 Forest [Green 100.00 98.62 787/787 787/798 Clouds [White 100.00 100.00 1077/1077 1077/1077 Cloud Shadow 95.68 98.94 930/972 930/940 Water [Blue] 96.84 94.38 705/728 705/747 Plains [Yello 99.59 100.00 727/730 727/727

Class Confusion Matrix Image Two Confusion Matrix: [Memory1] (1400x1400x1)

Overall Accuracy = (6366/6572) 96.8655% Kappa Coefficient = 0.9619

Ground Truth (Pixels) Class Rock Water Clouds Cloud Shadow Plains Unclassified 0 0 0 0 0 Rock [Maroon] 1206 0 0 2 0 Water [Blue] 0 970 0 105 0 Clouds [White 0 0 1535 0 0 Cloud Shadow 0 14 0 592 0 Plains [Yello 4 0 0 0 857 Forest [Green 1 0 0 0 57 Total 1211 984 1535 699 914

Ground Truth (Pixels) Class Forest Total Unclassified 0 0 Rock [Maroon] 0 1208 Water [Blue] 0 1075 Clouds [White 0 1535 Cloud Shadow 0 606 Plains [Yello 23 884 Forest [Green 1206 1264 Total 1229 6572

Ground Truth (Percent) Class Rock Water Clouds Cloud Shadow Plains Unclassified 0.00 0.00 0.00 0.00 0.00 Rock [Maroon] 99.59 0.00 0.00 0.29 0.00 Water [Blue] 0.00 98.58 0.00 15.02 0.00 Clouds [White 0.00 0.00 100.00 0.00 0.00 Cloud Shadow 0.00 1.42 0.00 84.69 0.00 Plains [Yello 0.33 0.00 0.00 0.00 93.76 Forest [Green 0.08 0.00 0.00 0.00 6.24 Total 100.00 100.00 100.00 100.00 100.00

Ground Truth (Percent) Class Forest Total Unclassified 0.00 0.00 Rock [Maroon] 0.00 18.38 Water [Blue] 0.00 16.36 Clouds [White 0.00 23.36 Cloud Shadow 0.00 9.22 Plains [Yello 1.87 13.45 Forest [Green 98.13 19.23 Total 100.00 100.00

Class Commission Omission Commission Omission (Percent) (Percent) (Pixels) (Pixels) Rock [Maroon] 0.17 0.41 2/1208 5/1211 Water [Blue] 9.77 1.42 105/1075 14/984 Clouds [White 0.00 0.00 0/1535 0/1535 Cloud Shadow 2.31 15.31 14/606 107/699 Plains [Yello 3.05 6.24 27/884 57/914 Forest [Green 4.59 1.87 58/1264 23/1229

Class Prod. Acc. User Acc. Prod. Acc. User Acc. (Percent) (Percent) (Pixels) (Pixels) Rock [Maroon] 99.59 99.83 1206/1211 1206/1208 Water [Blue] 98.58 90.23 970/984 970/1075 Clouds [White 100.00 100.00 1535/1535 1535/1535 Cloud Shadow 84.69 97.69 592/699 592/606 Plains [Yello 93.76 96.95 857/914 857/884 Forest [Green 98.13 95.41 1206/1229 1206/1264

Question 1: Describe which class has the lowest user accuracy. Is that accuracy still acceptable? Image one, water was the lowest at 94.38 Image two, water was the lowest at 90.23 Question 2: Compare the two classified images (time 1 and time 2) and state what you find. How do they compare? Images are very similar and the same features are easily located in both images e.g. mountains.  Question 3: Which classification is more accurate? Image two seems to be the more accurate of the two images. Image two more accurately classifies the water features and also the rock features, this could be due to the images lack of multiple clouds and multiple shadows. The image could also classify rock better due to the lack of vegetation found in image two.  Question 4: What is the largest changing class in km2 (note you can change the units under Options) and %? Look at the Class Changes. Clouds: 361.21   Question 5: Compare your results with the “change detection lab” from report 2. Discuss how the changed classes are visible in your change detection