Wykrywanie dróg z wykorzystaniem transformaty falkowej

W niniejszym artykule na wybranym polu testowym przeprowadzono próbę wyodrębnienia dróg z wykorzystaniem informacji o intensywności pochodzącej z danych laserowych oraz informacji geometrycznej i radiometrycznej zawartej w zdjęciach lotniczych. W procesie badawczym wykorzystano możliwości jakie niesie transformata falkowa w zakresie ekstrakcji krawędzi na obrazie. Końcowym etapem prac było scalenie informacji. Uzyskany obraz przebiegu dróg jest zgodny z istniejącą ortofotomapą.

Road detection based on remote sensing data is an important research theme allowing performing many spatial analysis. Recently the possibility of road network extraction using the elevation data from airborne laser scanning has aroused great interest. The carried out studies revealed the potential of LIDAR data, but also their limitations and shortcomings. Based on the performed investigations it can be concluded that the most reliable results are obtained by integrating the laser and image data. In this paper the road extraction test is carried out on the selected test field, using the information of the intensity derived from the laser data and the geometric and radiometric information contained in aerial photographs. Based on the laser data, the exact parameter of intensity, the initial image of the roads was generated with use of simple morphological operators. The next step was to detect the edges based on orthophotos using the wavelet transform. Wavelet transformation proved to be a useful tool to detect sudden changes in brightness. The final stage of the work was to merge the information from these two sources. The resulting image corresponds to the roads in the orthophotomap. In conclusion, it can be stated that the integration of photogrammetry and laser data is the optimal approach to solving problems of spatial object detection.