Design Of Recognition System Of Disordered Cigarette Based On Image Recognition

The coiling and packaging machinery used by tobacco manufacturing enterprises belongs to high-speed automatic production equipment.With the development of society and technology,the level of automatic detection on the roll-to-roll equipment is getting higher and higher.Today,most of the hood equipment also has an appearance imaging test,which can effectively control the quality of the package and the appearance of the small box.Since the messy cigarettes have little effect on the quality,there is no relevant research and corresponding detection so far;however,the messy cigarettes are one of the main reasons leading to the equipment stoppage.Seeking an effective method to detect the disorder in advance is the main research of this paper content.First of the all,there is the introduction of the principle of machine vision,this paper combines the cigarette production process to explain why the messy cigarettes will cause the equipment to stop,and analyze the adverse effects of equipment shutdown.According to the process flow,the requirements analysis of the identification system is proposed,and based on this,the design scheme of the identification system block diagram and the function of each module is determined.In this paper,the positioning block is used to register the image to facilitate the identification of messy cigarettes and correctly locate the cigarette tributaries.After correctly extracting the cigarette tributary,this paper lists three recognition algorithms:Hough transform to identify the target,identify the horizontal smoke color block target,and identify the horizontal smoke texture recognition target.And compare the advantages and disadvantages between various algorithms,and found an efficient and simple method to identify the target of horizontal smoke color block recognition.On this basis,considering the characteristics of the project,considering the hardware costand easy realization,this paper chooses the Raspberry Pi to build a messy cigarette detection system.Based on the correct connection of the hardware circuit diagram and the hardware such as the mounting rack,and the correct installation of the relevant plug-ins,the experimental platform is built on the basis of the previous system block diagram and the identified color blocks.Observe the recognition effect of the system in an hour.The final result shows that the identification of cigarettes in the front of the cigarette tributary is better,and the recognition rate is high.However,the recognition algorithm and recognition method for the messy cigarettes after the cigarette tributaries need to be improved.The project provides a feasible method for identifying messy cigarettes,has a strong practicality,and has important guiding significance and economic value.