الفهرس 
AcknowledgementOnly 14 pages are availabe for public view
 |  | from | 96 |  |  |
| | | from | 96 | | |
Abstract
Recently,motion detection is essential in computer vision applications because of the significant demands for developing digital video reordering systems (DVR).We present a new technique for motion detection, which can handle the challenges of motion Detection. The proposed technique can be used in real-time surveillance camera systems. It can capture the motion from any type of camera that is extracted from each enrolled images or video for each moved objects. We implemented the processing and dissemination stages for the processed images that are used for digital surveillance systems. On the other hand, the need for high recording quality with the increase of camera numbers required huge storage space.Therefore, we used a registration point in the case of traffic motions. In order to take advantage of storage space that made high benefits with four different processes of first obstacles of motion detection, the speed of the moving object, the presence of mobile cameras, moving background algorithm that responsible for Detecting movement A comparative study is presented in this paper to investigate the reliability and robustness of the proposed system.
There are two ways to discover movement: a) using the hardware. b) Using the software. Concerning using the software to discover motion, we can say that it is easy and cheaper than using the hardware. Due to the cameras recording quality, and an increase in the quantity of cameras and a demand for huge storage space, technicians resorted to the video recorder in order to take advantage of the storage space by which they got plenty of high benefits at first. Unfortunately, certain obstacles of motion detection appeared including the speed of the moving object, the presence of mobile cameras and moving backgrounds that may tend to interfere with the quality of motion detection, the presence of changing lighting, which makes the algorithm responsible for detecting movement that gives negative, inaccurate results or a shadow of the moving body, etc. Motion detection depends on comparing one frame to another and then it detects changes. The most important ways to detect movement are Background subtraction - temporal difference - Optical flow - template matching) Undoubtedly, Motion detection is considered one of the most important fields of computer vision and surveillance system in the world today.In order to calculate motion detection in a real-time, there are lots of techniques for live video. Most of these techniques are based on comparing the current video frame with the one from the previous frames or comparing one fixed background with the other frames.So, if there is an object in the beginning frame that disappears,the program will record all the next frames as a motion though we know it is not a real motion.To avoid the previous defect, we renew the first frame, but we will still have a problem with the static background.Our proposed framework devised an adaptive way to solve the previous challenges. One of the big challenges that affect motion detection is fast objects.Our proposed framework aims to present a new smart technique to detect moving frames in real-time applications; it requires only a little memory and computation based on motion detection in video frames. Experimental results show that this method can detect the moving objects efficiently and accurately through many digital video recording (DVR).The results of this study provide a new technique to overcome some challenge of motion detection. The result indicates the lowest of percentage of wrong classification (PWC).Border highlight achieved good result in recall, FPR, FNR, .F-measure, precision except the specificity so it is the best methods above all.Itmakes a frame to the moving object the benefit of this method can define the shape of the moving object also it flow the object.