Research On Monitoring System Of Remote ATC Navigation Equipment Based On Image Recognition

At present,the station management of the air traffic control navigation device relies on manual,and the image monitoring system relies on the remote camera to transmit back to the monitoring room through data collection and is monitored by the staff.This mode relies on labor,which is easy to cause fatigue for workers and cause hidden dangers in civil aviation operations.Therefore,this paper proposes a remote air traffic control navigation device monitoring system based on image recognition.The image recognition algorithm in the monitoring system is used to automatically obtain the instrument image information to obtain accurate meter readings.The air traffic control department becomes passive monitoring.Identifying security threats,turning post-mortem analysis of general monitoring systems into pre-warning and in-the-spot analysis,reducing false positives and false negatives caused by human factors,liberating operators from heavy monitoring and meter reading,and improving air traffic control navigation The level of technical support automation of equipment is of great significance.The specific research contents of the thesis are as follows:First,the needs analysis and design of the overall system architecture.the overall remote navigation system is composed of a power source,a camera,a processing equipment,an air traffic controller,and the like.The camera acquires the real-time air traffic controller image,transmits it to the processing equipment for preprocessing,ROI correction,character segmentation and recognition,and sends it to the supervisors to form a complete remote diagnosis system of ATC navigation equipment based on image recognition.The overall requirements of the navigation device panel display status remote diagnosis system were analyzed,and the whole system architecture was designed,including data acquisition,image preprocessing,ROI correction and character recognition sub-modules.Lay the foundation for overall technical research.Secondly,the corresponding panel image acquisition and preprocessing methods are studied for the characteristics of the navigation device panel.Firstly,the panel image acquisition system is introduced,then the preprocessing process of the panel image is discussed,including panel image filtering denoising,panel image graying,and equalization of gray panel image,After preprocessing,high quality panel image data that can be identified is obtained.Again,the ROI(Region of Interest)positioning and calibration of the navigation device panel image is studied.The selection of ROI is especially important for symbol recognition in panel images.This paper is based on morphology for ROI positioning.However,due to the existence of the original image tilt angle,it will affect the final character recognition accuracy.Therefore,this paper proposes an edge operator-based ROI correction method,which performs tilt correction on the ROI to lay the foundation for the final character recognition.Finally,the symbol recognition in panel image method based on composite classifier is studied.Firstly,the character segmentation and normalization in ROI are carried out,then the combination of statistical features and structural features is selected as the feature quantity.The template matching,support vector machine and back propagation neural network form the composite classifier to recognize the characters,thus achieving the final high.The purpose of precision character detection provides accurate data support for remote diagnosis of navigation devices.