| Camellia oleifera is the main economic forest in southern China.Hunan Camellia oleifera is rich in forest and fruit resources,and the development of the camellia industry ranks first in the country.In order to develop camellia related machinery and improve the quality of camellia fruit,the quality of camellia fruit needs to be tested Relevant research shows that the rate of fresh oil is an important criterion for evaluating the quality of camellia fruit,and there is an important correlation between the rate of fresh oil and the quality,shape and density of mature camellia fruit.It is time-consuming and labor-intensive to study the correlation simply by manual measurement In order to improve the efficiency of the statistical measurement of camellia fruit data,this paper designs a detection platform by studying the dynamic weighing technology of camellia fruit and real-time image size detection technology.In this way,data such as the weight and shape of Camellia oleifera fruit can be obtained intuitively and quickly The main work and conclusions of the paper are as follows(1)Designed and built the Camellia oleifera fruit detection platform and the control system of the upper computer and the lower computer,detailing the components of the system:conveying sorting module,weighing module,image vision module and fruit cup components.Using Labview as the platform,the software interface and program of the upper computer were developed,and the wiring circuit diagram and I/O module allocation table of the lower computer PLC were given.(2)Researched dynamic weighing signal processing technology.According to the structural characteristics of the weighing module,the whole process of generating dynamic weighing signals and the sources of interference are analyzed.The adaptive noise cancellation method based on the normalized LMS algorithm is studied.The FIR filter based on the NLMS algorithm is designed to eliminate noise interference.Through experimental analysis,the evaluation index of the filtering effect is established,and the order of the filter is determined to be M=48,and the iteration step factor μ=0.5 Calibrated by a symmetrical weight sensor,the fitting formula between the sensor output voltage and the true fruit weight of Camellia oleifera was obtained y=45.351*x-57.402;finally,the dynamic weighing experiment of Camellia oleifera fruit was carried out at a transmission speed of 0.33m/s.Experiments show that the method studied has a good suppression effect on the dynamic weighing signal noise.The relative average error of the weight value of camellia fruit is about 1.284%,and the weighing accuracy is high(3)The method of measuring the size of Camellia oleifera fruit is studied,including image preprocessing and size extraction.In terms of image preprocessing,the image region of interest(ROI)is extracted from the Camellia oleifera image,and the RGB color space is converted into the HSV color space to obtain the S component image;through experimental comparison,the median filter algorithm is used to perform the Camellia oleifera image Image denoising process,using the maximum inter-class variance(osut)algorithm to segment the camellia s-component map,and using the sobel algorithm to extract the edge contour of camellia oleifera;for the extraction of camellia fruit size,by determining the centroid of the camellia fruit image first,and then by The method of traversing the outline pixel points of Camellia oleifera fruit to get the fruit height and diameter of Camellia oleifera fruit;finally,the vision system is calibrated,and the size conversion relationship between the pixel size and the size of Camellia oleifera fruit is determined by repeated calibration experiments as y=0.526x.(4)Experimental verification of dynamic weighing and size measurement of Camellia oleifera fruit was carried out.The experimental results show that:in the dynamic weighing experiment of Camellia oleifera fruit,the maximumrelative error of weight is 3.81%and the minimum relative error is 0.47%;the relative average error of overall weighing is 1.284%;3.03%,the minimum relative error is 1.34%,the relative average error of overall fruit diameter measurement is 1.77%;the maximum relative error of fruit height is 2.43%,and the minimum relative error is 1.43%;the relative average error of overall fruit height measurement is 1.82%;Smaller error,high measurement accuracy,meet system requirements. |
