Ordering the imagery with right parameters is crucial for getting the best value for your investment.  Multi-spectral Imagery contains rich spectral content that can be exploited for variety of applications in upstream as well as down stream applications in the Oil & Gas industry, for encroachment monitoring for pipelines, extracting impervious surfaces for storm water billing and watershed modeling, change detection for environmental studies, vegetation analysis for agriculture and forestry, and others. When ordering imagery, users are often presented with myriad of options or not enough parameters to select the right imagery for their application. The three order parameters that will ensure you order the right imagery are: 1. Number of spectral bands 2. Radiometric Resolution (i.e. 8 bit or 16 bits) and 3. File format and tile size

 

Number of Spectral Bands:

Despite increasing global availability of remotely sensed imagery, from both aerial and satellite sources, imagery is primarily used as a pretty picture or a backdrop in GIS applications. Most of these multi-spectral sensors can collect information in the visible and Near Infrared (NIR) portions of the sun light. The information collected in the NIR, not visible to human eye, has valuable information that can be used to extract a variety of features such as roads, buildings, vegetation, water, and others. The following picture shows an example of an image in both Natural Color as well as Color-Infrared bands.

  

Notice how the vegetation in the CIR image is red in color and the tonal variations of red denote the health of the vegetation.  The brighter reds depict healthier vegetation and darker reds depict vegetation that is stressed.  Figure 2. shows some of the basic features that can be extracted from imagery.

 

Radiometric Resolution

Most of the digital sensors, aerial as well as satellites, can collect data at 11 to 16 bit resolution, which results in detailed information that can be exploited in the feature extraction process. Figure 3. is an example of detail in shadow areas of a 12 bit image which would have been lost if the customer ordered 8 bit imagery.

 

If the end user ordered 8 bit data, typically the data in the shadows and on bright objects such as roof-tops is lost.

 

File Format:

Ordering imagery in file formats such as GeoTIFF, ERDAS Imagine, PCI .pix format, that is not lossy, will preserve the spectral integrity of the scene and can be used in image analysis. This option will typically result in large file sizes as compared to lossy compression formats such as JPEG. While ordering, the user can specify tile sizes that will result in files that are manageable and can be displayed in common GIS SW. The user has the option to later convert the 16 bit imagery into 8 bit for display purposes by using compression schemes such as JPEG, ECW, and others to reduce the file size.

 

Recommendations:

The author recommends that when ordering data, ensure the image specifications include: 1. All  spectral bands the sensor can collect 2. Full radiometric resolution, typically 16 bit imagery, and 3. Imagery be delivered in a format such as GoeTIFF that is not lossy. For questions or comments regarding this article, please contact the author at knavulur@pixxures.com.