Indoor Navigation Assistant System
Nearly 285 million people are estimated to be visually impaired worldwide and 39 million are blind and 246 million have low vision. These visually impaired and also elderly people face difficulties in navigation due to obstacles in their path. GPS technology is available for outdoor navigation but it cannot be used for indoor navigation. Indoor Navigation Assistant System will assist visually impaired and elderly people to navigate safely in indoor environment. The system will detect objects in the users path in real time while navigating and will give voice response if obstacle is detected. Various object detection and tracking algorithms are used to solve the problem.
Keywords: Image Processing, Video Segmentation, Object tracking(Optical …show more content…
The system is real time so first you have to capture the video i.e the user will take help of some device or camera to record video. The framing of the video has to be done so as to calculate the distance between two objects using adjacent frames. Optical flow algorithm is used to calculate the distance between two points or pixels. Vibration elimination is used to remove any noise in the frames. Noise can occur due to camera movement. It is an important step to get accurate results. Next step is object detection and tracking. This is the final step or the important step in which you actually detect an obstacle. You have to make sure that you aware the user of the system that the object has been detected. So the voice response step comes in which you can tell the user as soon as obstacle has been …show more content…
After background image is obtained, background image is subtracted from the current frame. If the pixel difference is greater than the set threshold T, then determines that the pixels appear in the moving object, otherwise, as the background pixels.The moving object is detected after threshold operation. But background subtraction method is sensitive to the changes in the environment.
In temporal difference method the current frame is subtracted from the previous frame, and if the difference in values for a given pixel is greater than a threshold (T), the pixel is considered part of the foreground. In this method, detection of moving object is not accurate. This method is the simplest form of background subtraction.
Color histograms have been widely used for object detection and tracking due to their robustness, speed and simplicity. Color histograms are stable object features in the presence of occlusion and over changes in views, scales and shapes.