Computer-assisted photointerpretation aids to forest inventory mapping: Some possible approaches

Aerial photointerpretation is the mainstay of management forest inventories . Accuracy and consistency of this interpretation is a major concern. In ad dition there is a growing demand for these inventories to provide more quan titative information and data on new forest parameters. Incorporation of co mputer-assisted techniques to the interpretation process offers potential t o improve forest inventories along these fronts. Techniques must fit into the current infrastructure of mainly small or inde pendent interpreter groups or individuals using 1:10 000 to 1:20 000 scale stereo photography. Computer-assisted interpretation therefore must be base d on digitized photography and not necessitate large computers or costly di splay systems or analysis software. They must be simple to apply and not re quire inordinate fine tuning or trial and error by the interpreter. Envisaged techniques are: 1) computer-based interpretation keys, 2) recall acid display of ancillary data, 3) computer-assisted interpretation review where interpreters can quickly compare all stands they have interpreted as a given stand type, 4) automated single tree analysis including automated s ingle tree isolation, delineation and species classification, plus estimati ons of stems/ha, crown closure, crown size, numbers of snags, and gap or pa tch size distribution, and 5) pixel and area based features such as convent ional pixel classification and various texture measures for stands that cou ld be presented visually or numerically to the interpreter as an aid. As we ll, visual or parametric comparison of current and past inventory photograp hy presents a vast array of possible benefits to interpretation.