Hangar Safety Supervision System Based On Intelligent Image Processing

In view of the powerful features(such as clear, vivid and informative), Digital Video Supervisory System(DVSS) has been widely used in our daily life and work, and plays a more and more important role in many fields such as resource, national defense, traffic, etc.Intelligent DVSS can automatically analyze sequential images by the methods of computer vision and video analysis. When abnormality happens on the location monitored, the system alarms automatically and accidents could be avoided.With the rapid development of computer's capacity, network and image processing, the intelligent safety supervision system has become a new trend. Intelligent DVSS can significantly reduce staff labor intensity, improve work efficiency with less manpower, money and material resources, which has wide application prospects and important application value.This dissertation presents the design of a fairly intelligent hangar safety supervision system with practical background. On the basis of the existing real-time network control and network preview functions, the new hangar safety supervision system fulfills the function of anti-collision by using image processing technologies, such as object recognition and motion detection. It includes the following several topics: motionless target identification and allocation, detecting, extracting and tracking the moving objects under the complicated background. With the system specific applications, the method of image segmentation based on color features clustering is used to extract the features of motionless airplane. Here presents the improvement of frame difference: Gap-framed difference based on hue of the image. It used successfully in the detection of motion airplane.Using color features clustering technology, the former transfers the color images into binary images directly, then abstracts one of the objects' features.This method performs well in image segmenting in the condition of some complex background where illumination changes frequently. The later avoids error detecting effectively and enhances the speed greatly by using gap-framed difference, which makes full use of hue difference in different objects. It's a fast and accurate algorithm to detect and track moving objects. Because the two algorithms are practical and effective, they are greatly enhanced the image processing speed and efficiency, and is particularly applicable to real-time image processing.All the work was performed on Microsoft Windows XP operation system, and the algorithms were realized with C/C++ programming language on the Visual C++ 6.0 developing platform. Multithreading technology is used and the storage management is optimized to make the system become more stable.Based on the video data recorded from a hangar, experiments show that the system designed in this dissertation is stable enough. It runs in a good way to recognize motionless airplane and implement real-time motion detection and tracking of moving airplane, and it acts as a safeguard very well. All the results indicate that the system is adaptive and robust to the complication of background.