EVS-Islands

These stutter vectors occur throughout the world on the PGS data base. They are obviously a hiccup in the automated vector drawing tool used by EarthSat. They show up as you zoom in. Are they a problem? If NGS paid per vector they definitely were overcharged. If the user does not zoom to the 1:7,280 scale, they are barely visible. However, if you expect PGS to track the visible shoreline, they are a problem. They are a problem that occurs throughout the PGS data base. These stutter vectors should be cleaned up.

These sharp turn vectors are angles that do not track the coastline. Again they are found throughout the PGS data base, so it happened during the vector creation phase. You can follow a curving shoreline that periodically has these sharp turns interspersed throughout. Again, they are barely discernable at 1:62,500 and above. However, they are part of the PGS data base and they do not track the shoreline accurately. They should be corrected as well.

EVS will correct all of these issues. EVS, properly administered, identifies the visible shoreline, and then tracks it with greater precision than PGS. PGS took a great deal of time, effort and money. A team of volunteers needs to adjust PGS to EVS precision. A one-time effort by volunteers using rigorous guidelines and then having their work QC'd to ensure an EVS quality map file.

Posted by Mr Minton at 7/22/2006 12:18:00 PM | Permalink | Comments | links to this post

To do this post justice the reader should Left Click on the image and then select "All Sizes". You will be able to view the image at it's original size 1280 x 1024. The shorelines and map vectors are much easier to study.

This begins with the above split image. On the right is Landsat N-48-15 showing the area around Vinh, Vietnam. The image is at 1:250,000 or the same scale used by World Vector Shoreline (WVS). In my opinion, PGS shorelines are gorgeous at this scale. They are very precise.

I have been told by numerous people that work with Landsat ETM+ that this is the most reliable scale, 1:125,000, that can be extracted from Landsat. I know all about the limitations of this imagery (14.5 meters per pixel), but I use my eyes and it looks like one could zoom in even further. So lets try!

At this scale, 1:62,500, I still believe that the vector files that have been created from Landsat imagery are PGS files. They are consistently tightly following the shoreline (as defined by EarthSat algorithims). There are some clunky looking angles, but the overwhelming majority of vectors are close to where they should be.

At this scale the clunky vectors begin to appear with greater frequency. The smooth shoreline is not so smooth. But could PGS be more precise?

Now the clunky lines are occuring with greater frequency. The PGS vectors were never intended to be used at this level of resolution. The shorelines are intented to be used at that 1:125,000 or, at best, 1:62,500. PGS and WVS were never made to be used at this resolution. But, can these PGS vectors be improved upon? Yes! Landsat imagery begs to have the vectors that claim to track the shoreline actually track it.

At 1:7,810 scale the PGS vectors kind of track the imagery depiction of the shoreline. Then comes Enhanced Vector Shorelines (EVS).

The brown line is my EVS shoreline. The upper portion of the image tracks a sand bar that encircles a small lagoon. Many of the unneccessary vectors have been deleted. It is as if the digitizing algorithim stuttered. Almost all of the vectors have been adjusted. The adjustments are typically between 10 to 50 meters. Some might say that I need to back off. Just live with the 10 to 50 meter differences. But, I say Landsat depicts a shoreline and it is my job to determine it and then to place vectors that trace it. I call this effort EVS. Look at the final image (Select it, select "All Sizes" ) and you will see a shoreline that does justice to Landsat ETM+ 2000.

Posted by Mr Minton at 7/19/2006 08:38:00 PM | Permalink | Comments | links to this post

I do have a life beyond maps and map-making. Yesterday, I worked on computers in my classroom. Regular readers know that I am a middle school teacher in my other life. On my summer breaks I spend time in my classroom preparing it for the next school year. This summer, I am installing three donated Dell Optiplex Pentium III computers. I am not a computer nerd. I am a map nerd, not a computer type.

I have loaded them with operating systems, programs and even got them to talk successfully to my wireless router. The students will have three, Internet-ready computers to work with next school year.

The students will prepare PowerPoint presentations on people and events within World and US History. They also conduct Internet research on assigned projects. I use a laptop connected to a projector to teach the students history lessons. Every major unit is prefaced by a Google Earth Tour of the site(s) where things happened. For current events I use Google Earth and we tour the "hot-spots" throughout the world.

The students also work with maps. They routinely create political, physical, and thematic maps. Every chance I get to use a map to teach, I do it! I feel the students need to visualize history upon a map. I love that part of my job.

Finally, I will continue teaching a few students, in an after school program, the gentle art of map-making. I teach them how to make maps from the base image to the final product. It is great fun and the audience is always eager to learn. I use my experiences to teach them the methods to employ in order to generate a quality map. It is great fun.

So goes my fun summer. I also take my grandchildren for a swim as often as we can. It is a great way to beat the heat.

Posted by Mr Minton at 7/18/2006 07:58:00 AM | Permalink | Comments | links to this post

This is a continuation of my previous post, Namorik Atoll MH. As I stated, the majority of my digitizing came from the Google Earth image of Namorik Atoll. It is a 1-meter resolution image. As a mapping challenge, how detailed do you require? In the above map/image, I have added buildings. You could inspect the reef and accomplish detailed digitizing. 1-Meter imagery is found throughout Google Earth. The maps that can be developed from this level of imagery are extremely precise.

I must confess, the reefs in my previous post were digitized from Landsat-7 imagery. I did not require the level of resolution for reefs that 1-meter would afford.

Viewing this island, and others, at 1-meter resolution, one gains an intimate understanding of the layout and nuances of island life. It allows one to take a "poor man's journey". Happy travels!

Posted by Mr Minton at 7/17/2006 06:59:00 AM | Permalink | Comments | links to this post

Namorik Atoll (5˚35'N., 168˚07'E.), about 63 miles NW of Ebon Atoll, consists of two wooded islands on the reef enclosing the lagoon. A coral islet stands between them on the reef, with numerous black boulders. The very shallow lagoon is cut off from the sea by the drying coral reef. Boats can cross the reef with difficulty, at HW on the W side of the atoll. There is a trade's store on the W side of Namorik Islet, on the S side of the atoll.

Landing near the W side of Namorik Islet can be effected about 90m S of it. There is no shelter during NE winds, and it is dangerous with strong SW winds when there is a heavy sea. There is a fringing reef which extends about 135m in the vicinity of the landing place. There is depth of about 1.2m and there are rocks in places. Two stranded wrecks lie about 90m off the S shore along the reef line. (SD Pub-153)

Namorik Atoll has a population of 800+. It is small and out of the way. During WW-II there was fighting up at Kwagalen Atoll. According to sources Namorik Atoll required limited resources to retake it from the Japanese.

This map is another project derived from Google Earth imagery. They have many of the islands in the Pacific with 1-meter resolution. It is such a pleasure to work off of that level of resolution. The real power of this map comes when you zoom in to levels between 1:10000 to 1:5000. My digitizing of the shoreline was exceptional. This is a very solid effort.

Enjoy!

Labels: atoll, Digital Globe, EVS, Google Earth, island, Landsat, map, Marplot, Marshalls

Posted by Mr Minton at 7/16/2006 10:47:00 PM | Permalink | Comments | links to this post