In-depth anatomy towards analytic methods of image texture and analytic methods and study actuality of image texture analysis, combined with texture character of remote sensing image , the paper brings forward a new image recognition method of remote sensing which combines co-occurrence matrix with t test in statistics,and emphasizes on the theory, the algorithm, and the realization of the method in object recognition of remote sensing image.
The system which incoporates visualization analysis. raster GIS functions, and knowledge of mankind interpretation experts offers an effective way for solving such complicated problems as object recognition from remotely sensed images.
Based on an advanced hyperspectral Imager developed in this project and multiband and multipolarizotion radar imageries the imaging mechanism of rocks, vegetation, water etc, object recognition, information extraction and classification have been studied.
Based on the shape parameters and spectral library, a new algorithm has been carried to reduce the difference caused by atmospheric radiation, and recognize objects automatically. The shape parameter curve well describes the figure characters of objects, and more detailed information could be acquired by the new method. The algorithm has been applied for the object recognition of TM images,and the total accuracy reaches 88.96% in the test region.
The study conducted with the theories of remote sensing information in geosciences, hyperspectral and microwave remote sensing mechanism and object's recognition, remote sensing information transfer in different medium .
Therefore, it should be an important step in developing a system for automated perspective-independent object recognition.
Afterwards, a global similarity is calculated for object recognition.
When the numbers of templates vary from 4, 8, 18 to 36 for each object, and the remaining images compose the test sets, the object recognition rates are 95.75%, 99.30%, 100.0% and 100.0%, respectively.
The excellent recognition performance is much better than those of the other cited references, which indicates that our approach is well-suited for appearance-based object recognition.
Simulated quantum-optical object recognition from high-resolution images
At the last stage of the route, landmark recognition allowed them to pinpoint their preferred feeding site without using distant cues or odometric information.
By reducing the number of landmark recognition-triggered responses, this economical visuomotor strategy may be helpful in the Amazonian forest where many prominent landmarks are alike.
Five tests of spatial knowledge were employed: a route-length-estimation, landmark recognition, landmark ordering, map-drawing and navigation task.
These results suggest that route knowledge (landmark recognition and landmark ordering) requires less effortful processing than survey knowledge (developing a map-like representation and actual navigation).
First the approaches of landmark recognition and selflocalization for the wheelchair are proposed.