Developing And Testing A Combined Image Masking And Image Classification Procedure For Large Area Vegetation Mapping

Abstract

The Macaulay Land Use Research I n s t i t u t e Creigiehtickler ABFRllt-t N AB9 2Q3 UK ABS1 RAC1 Becausc o f ground r e s o l u t i o n e f f e c t s i t is o f t e n impossible t o map vegetation types d i r e c t l y from satellite imagcry. Howcvor, such imagery is o f t e n t h e only e v e i l a b l e source of v e g e t a t i o n information over l a r g e a r m 9 . I n order t o l l t i l i s e imagery i n l a r g e area vegetation s u r v e y s i t is therefore necessary t o develop i n d i r e c t mapping techniques which rely upon known associations belween plan1 d i s t r i b u t i o n and environmental factors, end only portly depend upon image-der ived i n rorniat ion. In t h i s paper t h e devalopmerk o f such an i n d i r e c t vegetst ion mapping procvc-lure is dcocribcd w i t h respect t o a n a t i o n a l s u r v e y o f bracken ( P t e r i d i t m s u i 1 inum,) i n Scotland. The f i r s t p a r t describes how direct rrtappiry techniqucs using supervised imagp classification methods on LANDSAT MSS imagery f e i l e d , and explains why i n tlie curltext or t h e i n t e r p l a y between t a r g e t si7e/stioye arid image r e s o l u t i o n . The second part d e a l s w i t h t h e development of an i n d i r e c t mapping procedure whj ch utilises stepwise image masking, based upon both e c o l o g i c a l and non-ecological decision rules, and msximum-likelihood classification o f t h e remaining image area, In the f i n a l section t h e results are presented and discussed i n t h e wider context o f e c o l n g i c a l p r o a b i l i t y mapping.