Note: This post represents the synthesis of the thoughts, procedures and experiences of others as represented in the 16 articles read in advance (see previous posts) and the discussion among the students and instructor during the Advanced Analytic Techniques class at Mercyhurst College on 15 April 2010 regarding Geospatial Analysis specifically. This technique was evaluated based on its overall validity, simplicity, flexibility and its ability to effectively use unstructured data.
Geospatial Analysis is a method using aerial imagery technology for analysis. These technologies include, Electro Optical, Infra Red, Multispectral, or Radar. This type of imagery is commercially available and useful for a multitude of applications. Analyzing geospatial imagery can provide decision makers with visual indicators that offer evidence of internal operations. Static images to aid the intelligence process by contributing to all-source analysis provided to decision makers. Imagery can be combined with cartography and mapping to determine distances, heights, demographics and population density.
- Objective in nature, data is structured.
- Easy to perform, even for untrained analysts.
- Readily available open sources such as Google Earth/ it is commercially available.
- Satellite imagery can provide data in remote areas that lack accurate geographic maps.
- Usually more cost efficient than sending personnel to a site.
- Can also provide powerful visualization to assist when shaping policy
- On the ground situation can be assessed, even when outside parties don't have access to the area.
- The same data can be used on multiple types of problems/issues.
- Technology is continually being updated to allow for easier use.
- Potentially expensive.
- Static images (only a snapshot in time).
- May require training depending on the level of analysis required.
- Errors may occur due to lack of resolution, deception efforts, or even weather.
- Imagery can generally be obtained at specific time intervals (i.e.- satellites cannot maintain stationary low earth orbit).
- Unable to view objects/structures located beneath the surface of the Earth.
- May require additional resources to provide specific measurements (e.g.-distance between objects, or heights of structures).
- Privacy and confidentiality issues may arise as technologies and resolution advance in sophistication.
There appear to be no standard procedures applicable to all images at all times or for all purposes. However, several standard processes for specific purposes were identified through the literature reviewed in advance of this class. For example, with respect to classifying types of vegetation:
- Images to be analyzed should show the same season (in case multiple images are used for analysis) and must be accurately registered (matched up) to the ground and to each other.
- The images must be radiometrically calibrated to minimize effects of instrument variations and atmospheric haze (Radiometric Normalization (RRN)). This is necessary when images are taken by different sensors, to standardize radiometry, or contrast within an image
- The landmarks in the images need to be aligned (geometric rectification). This is necessary for analysis over time.
- A classification scheme should be decided on and designed. Classification categories ex. cultivated land or grassland, forested land vs. non forested land can be used.
- Classify the images divide the photograph into "clusters", spatial areas with the same characteristics, then decide the category they should be placed.
- To reduce errors in classification, techniques such as having a human check portions of computer-classified work is also done (spatial re-classification).
- Classification results are compared to what the ground areas truly are (as revealed by aerial photographs) to assess the accuracy of the classification.
Geospatial Analysis, as a method, was demonstrated by looking at various images of major international cities. We had to identify them by viewing wide screen shots taken from Google Earth followed by viewing a close-up of the image. By recognizing key indicators, i.e. population density, landmarks, shadows, terrain, etc..., we were able to identify the cities (some, but not all) which taught us that as a method, it is best practiced using modifiers, particularly when resolution is lacking or distorted. The depth of Imagery analysis is dependent upon the source of the image (i.e. aerial vs. satellite) and the skill set of the analyst. This exercise, a descriptive analysis, would be an excellent starting point for those with low exposure to imagery analysis as it is low cost and easily accessible.