Detection of Afforestation, Reforestation and Deforestation (ARD) by Visual Photo Interpretation of High Spatial Resolution Images : A Fundamental Case Study

スポンサーリンク

概要

• 論文の詳細を見る

• Kyoto Protocol Article 3.3 calls for the identification of land use changes to better evaluate emissions and sinks of greenhouse gases. Here we investigated the ability of high spatial resolution imagery to examine changes in afforestation, reforestation and deforestation (ARD). Initially, we checked capability of the imagery in identifying land cover types by visual interpretation. Although digital orthophotos were capable of identifying various land-covers, a panchromatic HRV (HRV-P) image wasn't able to distinguish between young forests, crop fields and grasslands due to poorer picture quality in grey levels and spatial resolution than orthophotos. The poor quality of HRV-P images caused disagreement of ground object identification with the orthophotos by approximately 20-30% in the visual interpretation. Taking into the picture quality and the error causes into consideration, we visually interpreted ARD in Kumamoto prefecture using orthophotos for the beginning of the period and HRV-P images for the end of the period. A field validation showed that user's accuracy of detecting deforestation (D) was 92%, and that of detecting afforestation and reforestation (AR) was barely 50%. Although it was impossible to separate land-use changes from land-cover changes perfectly by visual interpretation, detecting deforestation was very accurate. The major causes of errors were interpreting cut-overs as D, and interpreting crop fields and grasslands as forest on the HRV-P images. Annual occurrence rates for AR and D were 0.001% and 0.048% of the land area, respectively. Annual occurrence of D was 0.032% of the land area according to a census. The interpreted D area was consistent with the census information. Thus ARD detection by visual photo interpretation, which clearly shows human-induced ARD areas, is suitable for meeting the monitoring requirements of the Kyoto Protocol. The interpretation using a Geographical Information System was better at identifying ARD areas because it used high resolution images with geographical coordinates, since the land-use change type and their locations were identifiable. We expect to be able to estimate total of land-use change areas with spatial information by decreasing the uncertainties in ARD visual interpretation and by introducing an efficient sampling method, which can estimate country-wide changes.

• 森林計画学会の論文

著者

• Awaya Yoshio

  Forestry And Forest Products Research Institute

• IEHARA Toshiro

  Forestry and Forest Products Research Institute

• Hayashi Masato

  Japan Forest Technology Association

• Hori Shuji

  Japan Forest Technology Association

• Matsubara Yoshitaka

  Japan Forest Technology Association

• Awaya Yoshio

  Gifu Univ. Gifu Jpn

• Awaya Yoshio

  River Basin Research Center Gifu University

• Awaya Yoshio

  Forestry And Forest Products Res. Inst. Ibaraki Jpn

• Iehara Toshiro

  Forest Management Division, Forestry and Forest Products Research Institute

関連論文

• VEGETATION CHANGE DETECTION USING LANDSAT DATA IN NORTHERN FINLAND : A TEST CASE USING SMOOTHING AND LAPLACIAN FILTERING (15th Symposium on Polar Biology)
• Aboveground biomass equations for individual trees of Cryptomeria japonica, Chamaecyparis obtusa and Larix kaempferi in Japan
• Detection of Afforestation, Reforestation and Deforestation (ARD) by Visual Photo Interpretation of High Spatial Resolution Images : A Fundamental Case Study
• Monitoring the Target Region of the Kyoto Protocol Article 3.3 with LANDSAT/TM Data
• A Photo Based Measurement System using a Measuring Camera
• Analysis of Canopy Structure in Beech (Fagus crenata) Secondary Forests using an Airborne Laser Scanner(Silvilaser)
• Accuracy Assessment of ARD Detection by RS
• Application of Automatic Binarization Method for Nationwide Forest Area Mapping using Satellite Imagery
• Methodology of Detecting ARD area by Remote Sensing
• Segmentation and Classification with Discriminant Analysis of QuickBird Multispectral and Panchromatic Data to Distinguish Cryptomeria japonica and Chamaecyparis obtusa Patches(Multipurpose Forest Management)
• Development of a Prediction Probability Model to Monitor ARD Areas
• A Semi-empirical Topographic Correction Method based on the Relation between Slope-aspect and Mean Radiance.
• Estimating Mean Height and Stand Volume in Broad Leaved Forest Stands using LiDAR(Silvilaser)
• Special Feature for Detecting Afforestation, Reforestation, Deforestation using Remote Sensing and Geographical Information.
• Estimating Diameter at Breast Height from Measurements of Illegally Logged Stumps in Cambodian Lowland Dry Evergreen Forest
• Practicalities of Non-Destructive Methodologies in Monitoring Anthropogenic Greenhouse Gas Emissions from Tropical Forests under the Influence of Human Intervention

もっと見る 閉じる

スポンサーリンク