Research On Visual Identification Method Of Concrete Particles In The Process Of Pneumatic Cleaning

As the development of highway construction in China progresses,the demand for road maintenance continues to increase.Road surface fragmentation is an essential aspect of road maintenance,and the currently used mechanical fragmentation methods still have many problems.High-pressure water jet fragmentation technology can effectively solve these problems,but the recycling of concrete debris after water jet fragmentation remains a critical factor restricting the development of this technology.Therefore,this paper investigates the treatment of concrete debris after water jet fragmentation using pneumatic cleaning methods,The identification method of concrete particles in the process of pneumatic cleaning is studied.The main contents of the paper are as follows:First,the experimental platform for pneumatic cleaning of concrete particles was systematically designed and built.According to the functions of the pneumatic cleaning experiment,the experimental platform was designed comprehensively,and hardware such as the camera,system control module,and light source type were selected based on practical needs.The experimental platform was then constructed,and the overall control process was explained.Second,research was conducted on single and multiple particle recognition and target feature information extraction.The single camera was calibrated using Zhang’s calibration method,and the camera’s various parameters were obtained,and the calibration results were validated for accuracy,which showed that the calibration accuracy was within 0.5mm.For single particle recognition,the white light source was found to have better recognition effects than blue,green,and red light sources under the auxiliary illumination system,and the maximum entropy method was determined to be the best segmentation method under auxiliary illumination.Morphological operations were performed on the segmented image to determine the particle target’s position and size.For multiple particles in different distribution states,such as discrete multiple particles with more background interference,special accumulation shapes formed by multiple particles,and accumulation angles formed by multiple particle stacking,filtering,edge detection,and angle algorithm detection were used to extract feature information conducive to cleaning and to plan nozzle positioning.Finally,a concrete particle cleaning control system was constructed based on visual recognition technology.The proportioning valve aperture,motor angle program writing,and data communication were implemented,and the blowing of particles with different particle sizes was completed.The accuracy of the algorithm was verified using encoders and testing instruments,and the results showed that the motor angle error was 4%,the pressure error was1.09%,and the flow rate error was 0.31%,meeting the requirements within the equipment’s achievable accuracy.