Tree feature extraction using image data obtained through virtual field server

  • Authors:

  • Xuefeng Wang;Masayuki Hirafuji;Xiaodong Li

  • Affiliations:

  • The Research Institute of Forest Resource Information Techniques, Chinese Academy of Forestry, CAF, Beijing 100091, China;Field Monitoring Research Team, NARC, Tsukuba 305-8666, Japan;The Research Institute of Forest Resource Information Techniques, Chinese Academy of Forestry, CAF, Beijing 100091, China

  • Venue:
  • Computers and Electronics in Agriculture
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

The application of field servers is proving to be increasingly crucial to the process of remote monitoring. These devices are built to continuously obtain large amounts of environmental and meteorological data and, at the same time, transmit back a vast quantity of in situ imagery. The question of how to more effectively utilize these data must be answered. This paper discusses the reconstruction of spatial information, as well as the collection of this information through technical methods. These actions are performed using computer vision based on field server imagery. In order to test and verify the technical approaches involved, such as calibration, matching, reconstruction, and so forth, images of the Xanthoceras sorbifolia Bunge tree were used. Two samples of X. sorbifolia seedling imagery were reconstructed. It was determined that the precision of the above results was satisfactory. These results demonstrate that the technical approaches can further extract deep information from images obtained through virtual field server. The calculation of image feature points for regular objects, in combination with affine geometry theory, can effectively shield image noise and lead to satisfactory results. Using the sum of the least squares dispersion, in combination with the epipolar line, one can reduce occurrences of image complexity (image matching that occurs during image reconstruction).