| In 2017,China's air conditioning output reached 180.4 million units.While improving people's living comfort,the energy consumption caused by air conditioning has become increasingly prominent.Improving the energy efficiency of air conditioning is of great significance for reducing energy consumption and coping with climate change.The valve core is a key component in the air conditioning valve to control the flow of refrigerant,and plays an important role in the control of the energy consumption of the air conditioner.The accuracy of the aperture machining accuracy needs to reach±1um.In order to avoid the production of defective products entering the market,it is necessary to test the aperture size of each valve core.At present,the company uses a pneumatic measuring instrument to measure the gas flux of the micropore of the valve core,and then judges the pass or fail by manual visual inspection,but even under the same criterion,due to the difference observation angle and working state of each worker,the observation results exists deviation,thus reducing the overall quality of the valve core.Combined with the defects of the current measurement methods and the inevitability of machine substitution,this thesis designs a machine vision based air conditioning valve core micropore online detection system.This thesis first expounds the difficulties and measurement methods of micropore high precision measurement,and proposes the feasibility of measuring the micropore with a pneumatic meter under a certain air pressure.Secondly,in combination with actual measurement indicators and requirements,the cameras,light sources and lenses for machine vision systems are selected and matched,and a complete online detection system is designed.The air conditioning valve core micropore online detection system mainly includes the valve core transport and measurement platform,the electronic control system hardware and software,the host computer software and the image processing algorithm.The electronic control system takes the STM32F103 MCU as the core,controls the stepping motor and the solenoid valve to perform the valve core grabbing,ventilation measurement,sorting and triggering the industrial camera to shot the readings of the barometer and the pneumatic meter.The host computer software receives and processes the image data,and communicates with the electronic control system through the serial port.Then the image processing algorithms is researched,mainly including the pneumatic meter buoy positioning algorithm and the barometer pointer positioning algorithm.In the pneumatic meter buoy positioning algorithm,the K-Means clustering algorithm is used to classify and locate the numbers,buoys and scales,and a fast digital identification method is proposed.On this basis,the corresponding values of each scale can be determined,and then determine the buoy interval.In the barometer pointer positioning algorithm,the Hough change detection line is improved to make it possible to quickly determine the position of the barometer pointer.Finally,the designed online detection system was tested,and four abnormal working scenarios were set up to fully test the performance of the whole system.The test results show that the system has high adaptability and stability,and can meet the long term work requirements.The average inspection time of each valve core is 2.4 seconds,which is faster than the detection speed of 2.5 seconds for skilled workers.In terms of detection accuracy,the false positive rate of good products is 0%,and the false positive rate of defective products is 1.47%. |
PDF Full Text Request