RESULTS

The Digital-Ortho Photo with a manual interpretation was the most accurate and used as the truth for this study. Manual interpretation of aerial photography requires many man-hours and is labor intensive. The cost of flying a mission, processing the data, and orthorectifying the image is high. Some of the highest quality commercially available digital-ortho photos are typically in the $200/mi� price range.

Airborne Multi-Spectral Digital imagery with an Erdas Imagine unsupervised classification is moderately accurate (62% as accurate as manual interpretation of the digital-ortho photo). Categorizing the classifications is only moderately labor intensive and any reasonably experienced remote sensor can separate the categories with minimal effort. The price of the data is dependant upon the nature of the mission, but the processing is not as intensive due to a lack of orthorectification.

Landsat MSS is not as accurate as other data due to the large pixel size (80 m) and the fact that the radiance values are averaged for the entire pixel. Labor required to analyze the data is minimal once the Tassel Capped algorithm is applied. Subjective analysis of pixel values and distribution is easily done. The cost of the data is very reasonable as the satellite is continuously in orbit and sending data back to earth on a continuous basis. The data is readily available from the U.S. Government.

CONCLUSIONS

• Accuracy and cost are directly related.

• The quality of the data is dependant upon pixel size (resolution).

• Manual interpretation is the most accurate and most time consuming.

• Choice of data type is dependant on budget constraints and required accuracy.

• AMDC data gives you "the most bang for the buck" by providing the most info (5 spectral bands), accurately (0.5 m resolution), at a moderate cost ($35/mi�)

HOME