Development Of Intelligent Detection And Control System For Brewing

Chinese liquor culture has a long history,but the traditional liquor-making technology belongs to the labor-intensive industry.High labor intensity and low production efficiency are its' feature,especially in the process of fermented grains steamer filling.The main work of this thesis is to study a steam detection algorithm based on image processing,and develop an intelligent detection and control system for the process of steamer filling.In this thesis,the knowledge of digital image processing technology and computer vision is introduced.The gray level processing,binarization,luminance normalization and filtering de-noising algorithms are used in image preprocessing to remove the influence of illumination and noise on steam identification.In order to obtain the spatial location of the steam point,we calibrated the camera to obtain its internal parameters.The extraction of steam characteristics.The first step is the detection of suspected areas of steam,because the steam has diffusion characteristics,the steam belongs to the moving objects in pictures.In this thesis,we compare the advantages and disadvantages between background subtraction method and frame difference method by experiment simulation,and finally determine the frame difference method to be the method of steam suspected area detection.Then we extract three characteristics of steam.These include color feature,LBP texture feature and shape feature,which are analyzed and extracted on the extracted suspected region.The color feature contains color moment feature and B-R difference feature.The characteristics of the steam were analyzed using the actual steam image,and the method of extracting each characteristic of the steam was introduced in detail.Then the BP neural network is designed to fuse the features of the extracted steam.We detailed analyze the design principles and selection of parameters.The experiment results show that the steam detection algorithm in this thesis can detect the position of the steam in the video accurately,real-timely and effectively in the process of steamer filling.At the end of the thesis,we design and implement a complete intelligent detection control system,including hardware design and selection,software system architecture,and applying the above algorithm based on enterprise demand.At present,the system has been developed and tested in a winery trial,greatly improving the production efficiency of liquor and reducing labor costs.