Research On Nondestructive Detection Technology For Mango Internal Quality Based On X-ray Computerized Tomography

Mango is one of the five most famous fruits in the world.Because of its rich nutrition,colorful appearance,delicate flesh and unique flavor,it is known as the "king of tropical fruits" and has high nutritional value and economic value.As the characteristic fruit of southern China,the main mango producing areas are distributed in Hainan,Yunnan,Guangdong,Guangxi and Fujian provinces,with a planting area of more than 1.7 million mu.In recent years,due to the long-term application of excessive chemical fertilizers and other reasons,the soil environment has continued to deteriorate.The occurrence of physiological diseases such as heart rot,hollow disease,gum seeds,sponge tissue and other diseases in mango fruit has increased year by year,seriously affecting The quality of mango.In some mango producing areas,the losses caused by these physiological diseases are as high as 30-50%.More importantly,although the fruits of these physiological diseases have lost their edible value and commercial value due to internal damage,the appearance looks intact,and there are no abnormalities compared with normal fruits.It is difficult to distinguish and detect by the naked eye.It presents new challenges to the quality inspection and grading techniques of mango fruit.In this project,mango fruit is selected as the research object,and the detection of internal quality characteristics such as pulp and core damage caused by common diseases,insect pests and mechanical damage is taken as the research content.The damage rate is used as the main indicator of mango quality detection.Tomography technology and image processing and analysis technology have conducted non-contact and non-destructive testing on the main internal quality characteristics of mangoes,and have carried out research on non-destructive testing and quantitative analysis of the internal quality of mangoes.Build a machine vision analysis system that uses non-destructive and automatic detection of the internal quality of mango fruit for industries such as fruit preservation,grading,processing,and export.The main work involves the research of image processing and segmentation algorithm of mango CT sequence,determining the location and size of the defect area,and the quantitative analysis method of mango damage based on image analysis.1.Through comprehensive analysis of existing mango nondestructive testing technology methods and domestic and foreign related agricultural product nondestructive testing technology,the research objectives,research content and technical route of this article are put forward based on the research needs of mango internal quality nondestructive testing technology.2.In order to obtain complete mango contours and suspected damaged areas,according to the characteristics of mango CT sequence images,research and exploration of suitable mango CT image processing and segmentation algorithms.The piecewise linear stretching method is used to enhance the contrast of the mango CT sequence image,and the filtered image is subjected to median filtering to denoise,and then the mango fruit in the CT slice is extracted from the background area by using area filling and difference image method.3.Accurate extraction of pulp defect areas is a key technology for internal quality inspection,which is directly related to the accuracy of defect detection.To this end,this paper studies the extraction and analysis algorithms of mango pulp defect areas.First,a series of processing such as binary threshold segmentation,area filling,difference shadowing,and morphological trimming are performed on the CT image after filtering enhancement to obtain a suspected damaged area.Because of the histogram analysis of the CT images of fruits,it is found that the parts with lower gray values in the mango are not all pulp defect tissues,but may also be the intra-nuclear voids formed by the shrinkage of the core.To this end,using the characteristics of the gray average value of the void area in the core is small,and the area of the connected area is much larger than the damaged area of the pulp,an appropriate algorithm is designed to eliminate the influence of this type of area.Finally,calculate the ratio of the area of defective tissue to the area of mango.4.In order to improve the detection speed of the internal quality of mango,a fast non-destructive detection method for mango damage rate based on CT image data of a specific number of layers is proposed.In the experiment,the mango sample image data used is a data selection scheme that takes one layer(n = 10,20,30,40,50)and a specific single layer from the CT sequence image data every n layers.After calculating the damage rate of the sample based on different compartments,through comparison with the detection method based on the overall CT sequence image data,it was found that 5 schemes such as 1layer every 10,20,30,40 and 50 layers.The detection times are 199.74 s,91.1s,60.06 s,49.78 s,40.37 s,which are 30.3%,13.82%,9.11%,7.55%,6.12% based on the overall CT sequence image data detection method;The relative errors of the rates were 2.24%,0.88%,6.17%,9.11%,and 13.69%,respectively.The comprehensive analysis results show that the detection technology based on a specific multi-layer CT image detection layer taking every 20 th layer has the best comprehensive performance,not only has good detection accuracy,but also the detection time limit is controlled below 100 s,effectively improving detection speed.