Results of Recognition and reconstruction

Standard Hough Transforms


Back

Non Noisy Image(Input):

Front View Side View
Top View

Output:


Front View(Before Post Processing)


Front View(After Post Processing)



Top View(Before Post Processing)


Top View(After Post Processing)



Side View(Before Post Processing)


Side View(After Post Processing)

Top


Noisy Images(Input):

Original Image1:




After Thresholding(Threshold =50) And Manually Removing Dimensioning:
Front View Side View

Output:


Front View(Before Post Processing)


Front View(After Post Processing)



Side View(Before Post Processing)


Side View(After Post Processing)



Original Image2:




After Thresholding(Threshold =50) And Manually Removing Dimensioning:



Image(Before Post Processing)


Image(After Post Processing)


Top


Reconstruction Of 3-D Wireframe model


We have generated test images for each case we have added in the previous algorithm which only dealt with lines. The reconstructed views are perspective.

Input1



Output1




Input2



Output2




Input3



Output3




Input4



Output4




Input5



Output5




Input6



Output6



Top




The Complete Algorithm



The above two sections show the results independently for recognition and reconstruction of the image respectively.This section shows the results for the complete program which we had made.It takes as arguments the three views as the input and efficiently recognizes and reconstructs the final image.

Non-Noisy Input

The Output

Noisy Input

The Output

Top   Back