Reconstructing And Watermarking Stereo Vision Systems

Stereo Vision is a very popular tool to recover 3D depth information from 2D flat images. Stereo Vision is used in many applications such as robot vision, aerial mapping, aerial navigation, 3D television, 3D cinema, 3D video applications and remote medical surgery. The reconstruction process has many challenging problems, for example the calibration problem, the processing speed and the correspondence problem. The correspondence problem occurs due to the mismatches which may arise while the matching is done between the left and right images. The pair of images is taken from different angles so they are not the same image; the illumination of the scene could affect the pair of images making them completely different. Also the occluded objects which may appear in one image and disappear into the other. All of these reasons lead to the mismatching problem which will directly affect on the resulting depth information.We have solved the correspondence problem by applying the matching process in two levels, the first level is Feature based matching, in which we have extracted the features of both images by creating multi-resolution images and applying histogram segmentation. The resulting features are region features; a comparison is done between the regions in the first image with the regions of the second image to get the disparity map.The second level is Area-based matching in which we applied the Wavelet transform to get an expected window size as a search area for each pixel. We have joined the two levels to obtain more accurate pixel by pixel correspondence. We also obtained an adaptive search range and window size for each pixel to reduce the mismatches. Our procedure introduced high accuracy results and denser depth information.The depth information is used to get the final 3D model– using only pair of images will create 2.5D model, using more than pair of images will create 3D model, we will refer to 3D model as a general output of stereo reconstruction– After reconstructing the model, in some applications it is needed to be published online. For example suppose the reconstructed model is a model for Sphinx– Famous statue in Egypt– The reconstruction for the model can be done in many days or months; then the model will be published online to let Internet users around the world watch the model. Therefore, techniques should be used to protect the copyright for that model. We have applied new fragile watermarking technique to secure the 3D reconstructed model and protect its copyright.Two common problems are usually encountered in 3D watermarking; the causality problem which arises due to the effect of embedding one pixel can propagate to the remaining polygons which constitute the 3D model, and the convergence problem which is the effect of embedding could lead to perceptible degradation for the model. In order to avoid the causality problem, we have proposed an efficient technique for traversing the polygonal meshes. The order of embedding is kept so the propagated effect of embedding is removed. Our procedure also solved the convergence problem by reducing the update of the vertices in the smooth areas and increasing it in the rough areas. The roughness of an area is calculated according to the angle between each pair of triangles in that area. The base edge of the triangle is then divided into small segments. The number of the segments will be increased in rough areas and decreased in the smooth areas, so that the vertex shift (move) will be imperceptible. We referred to our procedure as the Nearest Neighbor Move or NNM, because the vertex position will be moved to the nearest neighbor segment. The results of NNM show comparatively high visual quality for the watermarked meshes.We have used another type of watermarking technique to protect the copyright of the stereo reconstructed image. The procedure uses the obtained disparity map as a watermark. The Random LSB (Least Significant Bit) embedding technique is used to embed the watermark randomly into the left image of the stereo pair. The image is scrambled by applying Arnold transform before embedding the watermark. This technique not only protected the depth information from malicious attacks but also secured the watermark so that the attacker can't even realize the existence of a watermark.