Ice Profile Characteristics Analysis And Control System Based On Image Recognition

Rivers will form a closed ice sheet at low temperatures in the northern part of China.Due to the influence of human activities or natural factors,some areas of the river channel cannot be frozen.The exposed water surface in this area is prone to produce a large number of frazil ice under the effect of supercooling.The accumulation of frazil ice adhering to the bottom surface of the ice sheet will block the channel and cause the rise of water level,which eventually led to ice disasters.In response to this problem,relevant personnel have adopted a variety of methods to observe the thickness of the ice sheet and the particle density of the frazil ice,but it is limited by the test method and test accuracy.There are few reports on the automatic detection method of the thickness of the frazil ice layer.It is critical that the thickness of the ice sheet and the frazil ice be accurately measured to preventing ice disasters.In this paper,based on the color difference between the ice sheet layer and the frazil ice layer,a wavelet multi-scale edge detection algorithm and threshold adaptive Harris corner detection algorithm are proposed for accurate recognition of all aspects of ice image.In addition,for the actual needs of field applications,the image acquisition,remote transmission and motor drive parts of the system were further optimized,and a real-time monitoring system for the host computer based on the One NET cloud platform was built.The thesis mainly completes the following four aspects:(1)According to the characteristic that the ice sheet layer and the frazil ice layer have color difference on the image,the image recognition technology is used to process the ice sheet layer and the frazil ice layer image.By analyzing the shortcomings of classical edge detection and corner detection,a wavelet multiscale edges detection algorithm and the threshold adaptive Harris corner detection algorithm were proposed to distinguish the ice thickness.(2)The components are selected according to the field test environment,and the hardware circuit design of each functional module of the entire control system is completed,which mainly includes the STM32F767IGT6 minimum system circuit,image acquisition circuit,image storage circuit,remote transmission circuit and motor drive circuit.In consideration of the system power consumption problem,a relay switch circuit is designed.When the system stops working,the switch circuit powers off all peripheral devices to reduce power consumption,thereby ensuring long-term stable operation of the entire system.(3)Keil MDK programming software was used to design the programs needed for each functional module of the control system.MATLAB image processing tools were used to write wavelet multi-scale edge detection and threshold adaptive Harris corner detection programs.In order to remotely store the ice layer profile images and monitor the field test equipment,an One NET cloud platform real-time monitoring system was built.(4)According to the actual situation of the Yellow River,the ice environment was simulated,and different algorithms were used to process the entire ice profile.The threshold adaptive Harris corner detection algorithm(with the heating wire)was able to identify the optimal scheme of the frazil ice layer and its thickness.The field experiment is carried out on the river section,and the analysis results of various algorithms show that the wavelet multi-scale edge detection algorithm can be well applied to the detection of the ice sheet,and the maximum error with the artificial value is 2 mm.