The researcher have suggested real time depth based on frequency domain hole filling. It get better quality of depth sequence generated by sensor. This method is capable to produce high feature depth video which can be quite useful in improving the performance of various applications of Microsoft Kinect such as obstacle detection and avoidance, facial tracking, gesture recognition, pose estimation and skeletal. For stereo matching approach images depth extraction is the hybrid (Combination of Morphological Operation) mathematical algorithm. There are few step like color conversion, block matching, guided filtering, minimum disparity assignment design, mathematical perimeter, zero depth assignment, combination of hole filling and permutation of morphological operator and last nonlinear spatial filtering. Our algorithm is produce smooth, reliable, noise less and efficient depth map. The evaluation parameter such as Structure Similarity Index Map (SSIM), Peak Signal to Noise Ratio (PSNR) and Mean Square Error (MSE) measure the results for proportional analysis.
Real-time RGB-D image stitching using multiple Kinects for improved field of view