Use this URL to cite or link to this record in EThOS:
Title: Comparison of various soil survey techniques
Author: Albert-Ayolagha, Gaskin
ISNI:       0000 0001 3409 9457
Awarding Body: University of Aberdeen
Current Institution: University of Aberdeen
Date of Award: 1988
Availability of Full Text:
Access from EThOS:
Access from Institution:
A Field Survey and mapping was conducted after photo interpretation at a scale of 1:11,000 in an area of Northeast Scotland having varying geological features, physical conditions and land use. Mineralogical and Micromorphological analysis of modal pedo-units samples were carried out. Two Transects were marked out for land cover classification and soil mapping using the following remote sensing techniques: Aerial Photo-Interpretation (API), Landsat MSS and Airborne Thematic Mapper (ATM). MSS bands 4, 5, 7 and the corresponding ATM bands 3, 5, 6 were used in order to highlight the effects of spatial resolution. Five image classification techniques, Density Slicing (DS), Unsupervised Cluster Analysis (CL), Box Classification (BX), Maximum Likelihood (ML) and Visual Interpretation (VS) were used to separate six major land classes (urban, forest, agricultural land, moorland, bare soil and water). For soil survey and mapping the land cover classifications were compared with the various soil units in the study area. There is a weak correlation between the mineralogy and the soil types but the micromorphology correlated well with the soil types. Accuracy test for MSS August shows that ML has the highest classification accuracy having 70.0-75.6%. For MSS April computer aided VS having 72.1-77.6% is the most accurate. For the ATM the Unsupervised algorithm cluster analysis is the most effective having accuracy range of 86.6-90.6%. The comparison of the three remote sensing techniques shows that for land cover classification API is the most accurate having 93.3% followed by ATM bands (3, 5 and 6) CL having 86.6-90.1%. Landsat MSS had only 72.1-72.6%. The lower than expected value for the ATM is probably due to inappropriate and inadequate waveband selection (3 out of 11 bands). For soil survey the ATM CL analysis is the most effective. Accuracy for API and ATM are 70.6-76.2% and 86.9-90.9% respectively. It is expected that higher classification accuracy can be obtained from the ATM data by adequate waveband selection. Those suggested are bands 2, 3, 4, 5, 7, 10 and 11.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Soil Science & pedology