Research On Machine Vision Inspection Method Of Sea Cucumber Appearance Quality Based On Morphological Characteristics

With the significant increase in the proportion of sea cucumber dietary structure,consumers are increasingly high quality requirements for the sea cucumber.The appearance quality is an important standard to determine the commercial value of sea cucumber.At present,the detection of sea cucumber mainly has relied on manual evaluation.Artificial method has high labor intensity and strong subjectivity,seriously affecting the development level of sea cucumber food processing automation.In view,taking Bohai stichopus japonicus as the research object,this paper used machine vision technology to realize the classification of appearance quality of sea cucumber based on morphological characteristics.The main research contents were as follows:1.According to the design requirements of the machine vision system for detecting and grading,the paper completed the optimization of the system hardware,including the selection and build of black box,a light source,an industry camera,a lens and a computer.It designed and developed visual detection and grading system software based on MFC.2.Sea cucumber image processing method of segmentation,filtering,edge detection,position correction was studied in this paper.According to the experimental results and the requirements of classification,the paper selected the maximum between class variance method for image segmentation,5x5 template of median filter to remove noise and scanning method for boundary tracking edge extraction and correction on the location of the sea cucumber.3.For the extraction of sea cucumber characteristic parameters,the paper needed to reconstruct the body wall of sea cucumber.It used neighborhood comparison method and 16 neighborhood chain code method to extract features of sea cucumber body wall.It used parabola fitting to achieve the actual sea cucumber body wall restoration.According to the sea cucumber restoration integument and morphological algorithm,it made the morphology of sea cucumber burns to extract and analyse.it extracted and studied skeletons of sea cucumber axis by the Euclidean distance transform and constraint rules.4.According to the morphological characteristics of sea cucumber,the paper has selected six sea cucumber’s morphological parameters,such as perimeter,area,axial curvature,robust degree of burns,rectangularity and roundness parameters.Using artificial bee colony algorithm to optimize support vector machine parameters hierarchical model was established achieve sea cucumber exterior quality detection.Through training and testing,the model can be completed sea cucumber quality assessment.The experimental results show that the average recognition accuracy of the classification model for sea cucumber has reached 94%.