Research On Monitoring Technology For Aircraft Operation Based On Image Recognition

Airport is the basic operation facilities of civil aviation transportation system,which bears the takeoff, landing and ground activities of the aircraft. In recent years,the number of aircraft movements and airport operation in China's airport continue to increase, which extremely prone to congestion, then cause capacity bottleneck and security risks. To improve the operation efficiency and ensure the safe operation of the airport, it is necessary to effectively monitor the operation events of aircraft,such as takeoff, landing, sliding and others. At present, the computer vision technology is becoming more and more mature which be able to meet the requirements of the airport monitoring, and the camera price is cheap which can reduce the airport monitoring costs. In view of this, research on the theory and application of monitoring on airport aircraft operation is finished, which is based on image recognition technology, the main work and research contents are as follows:1. According to the characteristics of the airport monitoring scene, several moving target detection algorithms are analyzed. Adaptive background updating model is used to be algorithm for detecting moving objects in monitoring scene, and the background is reconstructed every fixed period to update the monitoring scene information.2. In the process of aircraft movement, the background difference is great, an improved adaptive threshold segmentation method is proposed, segmentation threshold is calculated in real time, morphological approach is used to fill target fracture and cavity.3. Taking into account that the same target have a small change in two adjacent frames, location, size and shape these three evaluation indicators are established, the evaluation function of each target between adjacent frames is calculated, target matching between frames is completed to follow multiple aircraft operation in the airport.4. An identification method for departure time of aircraft is presented. The runway surveillance area is defined in the image, two departure time identification model are respectively established that aircraft is from far to near and from near to the distant.5. An aircraft landing time identification method is put forward. HSV color model is used to split track area as an index matrix, and combined with the landing flight path of the aircraft to identify the landing time of aircraft.6. A risk identification method for aircraft during landing phase is proposed.Aircraft motion model is established, and kalman filtering is used to calculate the contrast between the normal flight path and detection flight path of aircraft.Monitoring video from Tianjin Binhai International Airport and Chaoyang navigation airport is used to verify, and the research results are effective. Auxiliary support can be provided for control officer to make policy decision.