Research On The Defect Detection Technology Of The Appearance Of Liquor Bottle

Images are involved in all aspects of the society.With the application of images more and more widely,image acquisition becomes increasingly important.Quality of the image acquisition system directly determines the quality of the samples,which will affect the final success of the entire process.That's why a good image acquisition system is particularly critical for the later image processing.At present,due to the urban transformation,especially the transformation of old residential buildings,the demand for intelligent acquisition of old gas meter is put forward,so it is urgent to develop a remote gas meter acquisition hardware system.Remote image acquisition of gas meter reading can be used to overcome the low efficiency of manual reading,and to effectively reduce the existing labor costs.Based on the remote image acquisition of a batch of old gas meters,this paper introduces the communication technologies such as the Narrow Band Internet of Things,designs an image acquisition system and discusses the relative algorithm of digital identification for the experimental samples.First of all,in this paper,according to the existing old gas meter,a Narrow-Band IOT wireless transmission mode of image acquisition system is designed.On the basis of the system the hardware has been upgraded to realize the transmission mode of NarrowBand IOT and 4G switchable,increase a wired transmission mode,expand the entire storage space of the system and provide hardware support for the realization of digital identification of the gas meter.Besides,the structure of the system is designed and the choice of light source is compared.Secondly,the hardware unit of the designed acquisition system was debugged,and the debugging results were analyzed to carry out the risk screening of the hardware function of the system.And the Internet of things platform was connected to multiple devices to realize the networking of multiple devices.Finally,this paper mainly designs the digital recognition algorithm for the experimental samples collected by the system.Through the image preprocessing such as filtering,binarization and threshold processing,the interference of the image is eliminated in the early stage.Roberts algorithm is mainly used as the edge detection of the Numbers in the sample,which is an important step to realize the template cutting of the digital model.The Numbers mentioned in this paper are identified by the 110 samples in the established sample database,and the corresponding experimental results are analyzed.