Research On Monitoring Method Of Runway Edge Lights Based On Image

The airport runway edge lights are mainly used to guide pilots to land safely and ensure that the light intensity meets the requirements,which is of great significance to ensure flight safety.Runway edge lights monitoring methods include manual inspection,airport flight aid light luminous detection vehicle and broken core detection and monitoring systems,but currently all these methods have disadvantages such as high cost or low efficiency.Therefore,an image-based monitoring method for runway edge lights is proposed.The status information and central light intensity of the runway edge lights can be obtained through processing the light image,which can efficiently and accurately monitor the runway edge lights online.Based on the analysis of the research status of the airport flight aid light monitoring system and detection methods at home and abroad,the following work has been done in this thesis:(1)Starting from field lighting images acquisition to demonstrate that the feasibility of setting image acquisition points on both sides of the runway,shooting light images remotely with the CCD camera and transmitting images with Zig Bee technology;then the overall system is analyzed from three aspects: image acquisition unit,Zig Bee networking communication and host computer software design.(2)The runway edge lights in the lighting image are classified into two categories according to whether they are qualified or not,and the deep learning target detection algorithm RetinaNet is selected to train and detect them,and output the light state information of the runway edge lights;the backbone network ResNet101 is optimized by network lightweight and the introduction of inverted residual module,and the anchor boxes suitable for the size of the runway side lights are found through cluster analysis,in this way,the accuracy of the improved detection model can reach 97.2%.(3)Through the gray scale conversion calibration experiment of the spot image and the modeling of the spatial direction light intensity of the runway edge lights,the method and calculation expression of the center light intensity predicted by the gray level of the runway side light spot image are determined.The central light intensity of the spot image of runway edge lights with 4-level light is predicted,and the error is about 13%.Based on the image of the runway edge lights monitoring method,with the light image processing,analysis and calculation as the main means,this thesis realizes the judgment of whether the runway edge lights are qualified and the calculation of the center light intensity.The experimental results show that the method in this paper can efficiently and accurately monitor the runway edge lights,and the error is within a reasonable range,which can provide a reference for follow-up research.