Submitted to: Remote Sensing of Environment
Publication Type: Peer reviewed journal
Publication Acceptance Date: 1/8/2006
Publication Date: 1/16/2006
Citation: Miao, X., Gong, P., Swope, S.M., Pu, R., Carruthers, R.I., Anderson, G.L. 2006. Estimation of yellow starthistle cover through casi hyperspectral imagery using linear spectral mixture models. Remote Sensing of Environment. 101:329-341. Interpretive Summary: Yellow starthistle is one of the worst invasive weeds to infest California and other western states. It grows in rangelands, pastures, natural areas, along roadsides and other disturbed habitats, infesting approximately 15 million acres. This weed produces toxins that cause brain lesion and death in horses, is a poor forage plant for other wildlife and livestock due to its spiny flower heads, it uses excessive amounts of valuable water out competing native species and other beneficial plants, and induces wildfires. USDA and UC Berkeley have worked closely to develop new effective remote sensing technology that allows aerial recognizance to more precisely estimate the area infested along with the percent cover of yellow starthistle across wide areas. This paper outlines methods of quantifying yellow starthistle within individual pixels collected using aerial borne hyperspectral imagers. Determination of these distribution patterns and infestation level are extremely important as they help land mangers develop and implement watershed level methods of managing this noxious weed.
Technical Abstract: The invasive weed Yellow Starthistle (Centaurea solstitialis) has infested between 10-15 million acres in California. It often forms dense infestations and rapidly depletes soil moisture, thus further preventing the establishment of other species. Precise assessment of its canopy cover, especially the low-density abundance in the earlier growing season, is the key to the effective management. CASI hyperspectral imagery was acquired at the western edge of California's Central Valley grasslands on June 15, 2003. Four linear spectral mixture models (LSMM) were investigated for original CASI data. Band selections based on residual analysis and feature extraction (PCA and LDA) were further explored to reduce the data dimension. The estimated percent cover of yellow starthistle was statistically adjusted according to the ground cover data. PCA-based fully constrained LSMM obtained almost the same results as that derived from the original CASI data, but largely reduced the computation burden. The uncertainty of PCA-based LSMM was estimated through Monto-Carlo simulation, and the maximum standard deviation was about 11%. The results suggested that unmixing CASI imagery could be used for estimating and mapping yellow starthistle for larger regional areas.