Research On Fruit Size-adaptive Packaging Control System Based On Machine Vision

At present,the packaging technology has been widely used in the industrial field and has become an important symbol of industrial modernization.However,the packaging machinery at this stage still has the characteristics of poor production flexibility,low packaging precision,and poor real-time response.The appearance of machine vision has pushed the development of the packaging industry to the intelligent stage.It has the characteristics of fast image acquisition,high detection accuracy,and wide detection range.The application of machine vision to the packaging industry can enable the packaging system to automatically complete the identification and detection of the packaged goods,and improve the packaging accuracy and the adaptability of the packaging system to the external environment.The thesis takes the control system of multi servo pillow packing machine as the research object,and studies the fruit packing control system based on machine vision.The original packaging control system has simple function and can only operate the equal-length package for fruit.then study the multi axis cooperative control strategy of machine vision image processing and packaging system,use visual system to obtain images,and use the processed images to determine the length of packaging bags,and then determine the multi axis speed cooperation according to the packaging parameters.The contents and conclusions of the specific research are as follow:1)The image processing method was studied.By comparing the processing effects of different methodS,the image processing algorithm in the visual system is improved.The fuzzy length is calculated by estimating the point spread function(PSF)of the blurred image,and the Lucy-Richardson algorithm based on gain control was used to restore the blurred fruit image.When the image iteration times r=8,the image segmentation error is minimized.After the fuzzy image restoration processing completed,the threshold method was used to conduct the segmentation of the fruit and the background,and then the image is subjected to edge detection and size extraction.The image size after the convolution operation was converted into the actual size by the camera calibration,and the error value between the image size and the actual size is eradicated,and the length of the packaging bag is determined by using the parabola method instead of the fruit circular line.2)Analyzed the packaging machine control system,and determined the corresponding packaging control strategy according to the variable-size packaging.After conducting model analysis,speed control was performed on the three axes of the system,and the variable-size packaging was realized by real-time adjustment of the delivery shaft velocity and the end-sealing tangent shaft velocity,which is to achieve a uniform speed of the film feeding motor.Solved the pulse frequency formula of its acceleration curve through the packaging demand.Changed the amount of pulses delivered to the fruit shaft and the end-cutting tangential axis to achieve seamless and accurate packaging between fruits.The initial pulse frequency f,acceleration a and pulse amount N in the acceleration formula are enclosed in the PLC envelope table,the envelope table data is updated in real time,and the speed control of each axis is realized by the PLC ladder diagram.Finally,the Queue function in the upper computer software Labview is used to implement the parallel processing of the image processing module and the motion control module.3)Hardware configuration of the system was completed through system components selection,including machine vision module and motion control module.The overall layout of the system was schemed and the data processing of the upper and lower computers was studied.The human interface was designed according to the system requirements.In the experiment,the algorithm was used to deblur the dynamic fruit image and analyze the results of the comparison process.The results showed that compared with the original unprocessed blurred image,the average measurement error of fruit maximum fruit diameter was reduced from 5.1%to 0.81%when using the proposed algorithm.And when the packaging velocity lower than 0.20m/s,the fruit packaging error was within an acceptable range.It proves the feasibility of the algorithm and control strategy chosen in this thesis,and provided the basis for the improvement of the entire control system.