As the resolution for multispectral images continuously improving, the phenomena that "synonyms spectrum", "foreign body with spectrum" become more apparently, therefore, in the results for pixel-based iteratively reweighted multivariate alteration detection (IR -MAD) may appear many issues such as broken patch, much pseudo-change, more noise, the low overall detection rate et al. In order to improving the above problems. In this paper, the pixel-based IR-MAD algorithm can be transferred to the object domain. the object-based IR-MAD(OB_IRMAD) method apply different and meaningful combinations of features rather than the original pixels. these features are classified into meaningful combinations, so that the results for object-based change detection can get higher reliability and accuracy. To stabilize solutions to the IR-MAD problem, some of regularization may be needed. A case with ZY-3 multispectral image at one point in time in province of Xinjiang border port demonstrate the effectiveness and feasibility of the OB_IRMAD. Compared as, using the same date and region we do the pixel-level IR-MAD change detection and artificial visual change detection. Finally, we calculate the various evaluation indexes for accuracy utilizing confuse matrix and compare the accuracy of the two detection results. The results show: in the ways of overall accuracy, correct rate, error rate, the OB_IRMAD is better than pixel-level IRMAD, change polygon more rules and indicates a less noisy.
In order to improve land cover classification accuracy of the coastal tidal wetland area in Dafeng, this paper take advantage of hyper-spectral remote sensing image with high spatial resolution airborne Lidar data. The introduction of feature extraction, band selection and nDSM models to reduce the dimension of the original image. After segmentation process that combining FNEA segmentation with spectral differences segmentation method, the paper finalize the study area through the establishment of the rule set classification of land cover classification. The results show that the proposed classification for land cover classification accuracy has improved significantly, including housing, shadow, water, vegetation classification of high precision. That is to say that the method can meet the needs of land cover classification of the coastal tidal wetland area in Dafeng. This innovation is the introduction of principal component analysis, and the use of characteristic index, shape and characteristics of various types of data extraction nDSM feature to improve the accuracy and speed of land cover classification.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.