ADIPRS Software
Advanced Digital Image Processor for Remote Sensing
By Dr. Eng.
Figure 1: Splash Screen.
Figure 2: Main Window Screen.
Figure 3: Image Processor Screen.
Figure 4: New Screen.
Figure 5: Edit Screen.
Figure 6: Selection Screen.
Figure 7: Selection Example.
Figure 9: Adjust Screen.
Figure 10: Histogram Processing Screen.
Figure 11: Filter Definition Screen.
Figure 13: Filter Color Screen.
Figure 16: Filter Special Screen.
Figure 18: Filter Special Map Grid Example.
-View Menu: It concerning with any option affects the display. Zoom option can affect the size of image display. All other options can hide or show tools in the module as shown in figure 19.
-Window Menu: As any other software.
-Help Menu: As any other software.
-Classification Menu: It is the main in this software package. It includes 2 types of projects (LandSat 7 project and general project) as shown in figure 20.
Figure 20: Classification Screen.
After choosing LandSat 7 project figure 21 appears. The user must enter 8 bands (1 and 6 are optionally while rest 6 bands necessary). Then the user can create new project and successful message appears if no errors happen. The user can use any shown buttons but here we will choose data processing so figure 22 appears. The user can compute covariance matrix of the data he entered. Also the correlation matrix can be computed. After that the user may decide to apply principal components transformation technique.
Figure 22: Data Processing Screen.
Returning to figure 21, the user can choose unsupervised classification and figure 23 appears. It shows that the user can choose the type of start of classification process. Also he can choose 10 supported classification methods (until now), 3 methods are Fuzzy-Based (HCM-AO, FCM-AO and PCM-AO) and 4 methods are Neural Networks-Based (GCLNN, CLNN, CLNN-LVQ and CLBPNN). And 3 are Neuro Fuzzy- Based (CLNN-FCM-AO, CLNN-HCM-AO and CLNN-PCM-AO). The user must enter names and desired display colors for his classes. The options choice will be varied according to the method as shown in figures 24, 25, 26, 27 and 28. The user now is ready to apply the classification. When the user presses the start button the iteration and pixels progress bars increased until the solution is reached. Then the matrix of clusters' centers appears with clusters values in each band and also the number of pixels in each cluster. The user can display the result of the classification process by pressing apply and choose the base image to be colorized as shown in figure 29.
Figure 24: CLBPNN Options.
Figure 25: CLNN Options.
Figure 26: CLNN-CMeans Options.
Figure 27: CLNN-LVQ Options.
Figure 29: Classified image display.
Figure 31: part 1 of classification report.
Figure 32: part 2 of classification report.
4 image can be displayed reference, classified, match area and no match area. Both user and producer accuracy computed in addition the kappa (khat) value as shown in figure 32.
