Research And Construction Of Smart Recovery System Based On The Internet

With China's economic development and urbanization,the amount of garbage in the city is increasing,bottles are an important part in which.In this case,the beverage bottle recycling plant comes into being.The intelligent recycling machines on the market today can only depend on unbroken beverage bottle barcode to recognize.The bottle cannot be recognized if the barcode is broken.Moreover,the barcode is too easy to forge,the recognition mechanism is required to improve.Secondly,the beverage bottle intelligent recovery device is only to collect bottles together and further classification is not made,which is so hard for beverage bottles for later processing.In this paper,the classification algorithm in machine vision is applied to the recognition and classification of beverage bottles,which is of great importance for beverage bottle inspection and further processing.In this paper,an intelligent self-recovery system for waste beverage bottles is designed.In order to identify the beverage bottle more accurately,the radio frequency identification device is used to identify the beverage bottle for the first time,then we propose the algorithm of object recognition based on bag-of-visual-words to identify the beverage bottle for the second time.According to the material the bottles are divided into plastic bottles and metal bottles and are put into their respective box.In this paper,we firstly present the background and significance of the subject and summarizes the development of machine vision detection technology and the application of image classification technology.Secondly,the function and specific work processes of the system are summarized,and the hardware structure of the system is designed according to this function.Then,some theoretical knowledge of image classification is introduced,and the concrete realization steps of applying the Bag of Words model algorithm to the recognition and classification of beverage bottles are introduced in detail.Finally,we debug the system and the performance measurement results of the model evaluation are given,which prove the effectiveness of the algorithm.