Information Extraction Of Euphorbia Distribution Based On Multi-source Image 祁连瑞 Incense

The alpine meadow of Qilian Mountains in Qinghai province with wide distribution and large area, which is an important component of the grassland ecological system. Its yield, productivity is of great significance for grassland ecosystem sustainable development and circulation. In recent years, due to overgrazing, and poisonous weeds invasion, especially the Stellera chamaejasme, because of its vigorous growth, resulting in the emergence of "poisonous weeds" phenomenon, resulting in the decline and serious degradation in coverage of alpine meadow.It has serious impact on their income and economic development and even threatens the harmony and stability of local areas. Using the remote sensing technology, giving rapid determination of poisonous weeds distribution range can provide scheme for the management of the development of eliminating poisonous weeds control which plays an important role in protecting the stability of grassland ecosystem.Based on the field measured spectral data, analyzing spectral differences of Stellera chamaejasme and its mainly forage characteristics,using spectral angle mapping classification algorithm for Stellera chamaejasme, information recognition. The main research has these work:(1) Analysis of Stellera chamaejasme and its mainly forage spectral features:the field measured spectral data of typical experimentation area Stellera chamaejasme and main forage, using the first derivative and red edge parameters such as spectral analysis technology, white flowers, flower buds of Stellera chamaejasme, dried flowers and leaf spectral characteristics, Stellera chamaejasme and mainly forage spectral differences, Stellera chamaejasme-different coverage community grass-soil mixed spectral features are analyzed. The results show that: in the blue band400-450nm, red band600-650nm and in the near-infrared band of760-900nm, Stellera chamaejasme is with diagnostic feature.(2) Classification of Stellera chamaejasme spectral mapping algorithm based on:the establishment of Stellera chamaejasme L. white, four main types of grass, grass-soil mixing, coverage and low coverage of Stellera chamaejasme spectral library. Elucidated by spectral library for endmember spectra, the spectral mapping classification algorithm for Rapideye and IKONOS images, of Stellera chamaejasme distribution information extractioa Through the classification of the classification results after treatment the categories and merge operations, the Stellera chamaejasme white and coverage,low coverage of Stellera chamaejasme recognition results classified as Stellera chamaejasme; the four main types of pasture and forage and soil mixing spectral recognition results classified as forage. The final classification is Rapideye and IKONOS images were divided into two categories of Stellera chamaejasme and grass, and making classification chart and compare.(3) The accuracy of the classification results of IKONOS/RapidEye image for Stellera chamaejasme verification:IKONOS and RapidEye images of Stellera chamaejasme area shows that:Rapideye image extraction area is consistent with the actual situation, the recognition effect is better than that of IKONOS image. Investigation shows that, for IKONOS images of Stellera chamaejasme distribution area and non Stellera chamaejasme distribution area of on-the-spot investigation results were significantly lower than that of Rapidyeye image, divided, misclassification and mixed phenomenon exists. While the Rapideye image recognition effect is good and is consistent with the root distribution in prospecting area with high recognition accuracy.