05/11/2011

Lesson 4 : GIS basics

Electric cables> line; District> polygon; Police and electric power stations> points
Lesson 4 : GIS basics

Having given tips how to install QGIS in my earlier blogs, I now get feedbacks probably from students asking how to use QGIS. In any course, there are 3 levels: basic, intermediate and advance. I think it is inappropriate to train intermediate and advance levels over the blog because it is difficult to put some explanation into words so I will limit it to GIS basic. My way of teaching may be a bit strange compared to a formal structured way at universities ( I was never good at that). I explain how to avoid getting lost in GIS as I did in the first place. In GIS basics, firstly know there are two GIS format ie. vector, raster then there is map projection. If you are into vector GIS features just know there are 3 i.e. point, line & polygon. We begin ironically by forgetting about GIS for a while to get a better understanding of what's it all about:


Firstly, we must realize the real world is 3 dimensional (3D). This means there are elements of length x breadth x height to consider when wanting to digitize a spatial form. Current GIS technology is way behind real 3D digital technology (to me, like a 3D movie) and can at best only translate the real world into 2D. I think if current GIS could really translate the real world situation which is 3D, the moment you want to create a GIS layer that GIS software it will force you into a cube-like dimension imposing on you to add x & y & z factor and even if the z factor was not available the view should be in a cube-like dimension if zoomed out. For now, the height factor in some "3D" applications is captured to produce an image that somewhat try to portray the real world but this is not really 3D in the true sense of the word. Don't let people fool you otherwise. Every point on this earth is translated by a coordinate reference point.


I explained how things get mapped in GIS well under the FAQs so I don't want to repeat them here. Look for this illustration below in the FAQs.

I have to do another drawing here because that flat plane is really a big vertical cylinder
the same circumference of the world and I am thinking how to best draw this without
giving the wrong impression... Mmm, still thinking here.
Simply put, if something physical has a geographical context then a minimum of 2 coordinate reference points are needed to illustrate that line. If height factor comes into play, current GIS technology not does not recognize it as a feature but as an attribute of that feature. Just remember the 3 GIS vector elements: point, line and polygon. To illustrate, a point can represent a lamp post or tree, a line can represent a road or rail line and a polygon can represent a plot of land. The physical entity will be a spatial drawing saved in the representative feature layer and they point, line and polygon maps do not mix so to see them all at once, the technique of sieving or overlaying is required. The attributes for that physical entity are kept in a corresponding attribute table kept hidden unless activated or vice versa.

Secondly, once a physical feature has been properly mapped onto a layer or what I would term as a digital paper, it is vital to set the proper map projection for that layer and if not precisely overlaid for the same locality, you can forget about doing GIS analysis. A layer without a map projection is a map floating somewhere in mid-air (don't ask me how high) yet precisely over a certain locality, it is not firmly set on the ground. They call this on-the-fly projection. To get the height element into proper perspective, that layer must come to the ground making height measurement taken from the ground. A map layer not properly projected then set to off-the-fly projection will immediately fly way off course from its original position almost half away round the world.


It is also important to realize in GIS, the round (more or else) world has been cleverly  flatten out by Mr. Mercator so unlike the real world where the North point is a point on top of the world, on a flat drawing, the North pole is a line somewhere "up North" so whether in America or in Malaysia yet on a flat drawing it is always perpendicularly up.


Thirdly, without getting too technical, GIS format can take the form of vector or raster format. Just know that a vector map is based on the creation of lines as illustrated in a cadastral map or landuse map. A raster map is made up of pixels like in a TV. Raster images are all solid/opaque just like a photography, aerial photo or satellite image.

Lastly before I forget, in computer aided designed (CAD) drawings, the space in a polygon drawing means nothing just like a rubber-band but in GIS that space has a spatial context and can be translated to mean a spatial feature e.g. landuse and because it has an area, it is possible to calculate the size of that space.

The translation of real world by raster format transforms images into a grid of cells or grid of pixels like we see in a TV, the more pixels in an image, the sharper the image becomes. To not get the image distorted, pixels are in square shape. Satellite images are translated onto a grid of pixels in different bands of light and because light reflects differently on objects and plants, different shade of light are achieved. Remote sensing agencies have graded and catalogue the different shades and can intrepret well the type of geographical features. Different types of crop cover show up well but different type of landuse activities do not though can be grouped as built environment.

Different specialization require different approach to meet their needs and in the field of town planning and land administration, vector gis is more preferred. Traditionally, people were use to the ESRI Simple feature vector model that separates the 3 basic vector feature point, line and polygon on separate layers.  When polygons shared the same boundary, it actually meant 2 separate boundaries. This meant the spatial area in the polygon could be recognized as the area. However, as GIS ideology and GIS programming developed, the GRASS vector model took a different approach. It is arguably closer to the real world situation where the 3 vector features of point, line and polygon are placed on a single layer (Don't ask me how they do it) just like every geographical object can be found on a flat piece or rolling land. Secondly, when polygons share the same boundary,the polygons do not need to close because the separate boundaries merge as one. This means that the area cannot be a close polygon because in certain situation, it does not occur. Thus to overcome this, the centroid of each polygon is created and there lies the data on the area of the polygon. GRASS vector  model thus stores attribute data in a slighty different way than the ESRI simple vector model. I don't want to go further at this point of time because I'm afraid, I will confuse you more so I'm leaving it up to here.

Well, that's so much about GIS basic. Did I get you all confused? :-D

7 comments:

  1. Sangat menarik dan mudah faham ! :)

    ReplyDelete
  2. post yg banyak membantu!!.. terima kasih banyak!.. can i share dis post?

    ReplyDelete
    Replies
    1. Of course you can! Tujuannya untuk berkongsi ilmu lah.

      Delete
  3. Replies
    1. Thank you very much. I always wonder when I get compliments like yours how is it that you can see what I want you to see but there are others who cannot see what you can see so where did I go wrong? :-o

      Delete