基于力学和图像特征的盐渍海参品质分级方法研究

Research on Quality Grading Method of Salted Sea Cucumber Quality Based on Mechanical and Image Features

海参含海盐量过高,对其食味品质和营养价值影响较大,为实现快速、无损盐渍海参品质分级,以两种不同食盐浓度下制备的盐渍海参为研究对象,利用质构仪采集力学数据信息,通过工业相机采集下压过程的图像信息,提取下压力作功的力学特征和目标形态变化的图像特征,分别将力学值、作功值、图像特征作为输入,选择最近邻分类器(KNN),构建盐渍海参品质分级模型,分级准确率分别可达77.22%、88.8%和88.8%。为提高分级准确率,将上述提取的数值特征进行归一化,以此为输入,建立基于KNN的盐渍海参品质分级模型,分级准确率可达到94.44%。该方法在盐渍海参品质分级应用中具有一定潜力。

The high salt content of salted sea cucumber has a great influence on its taste quality and nutritional value. In order to realize the rapid and nondestructive classification of salted sea cucumber, the salted sea cucumber prepared at two different concentrations of salt was studied. Mechanical information was obtained by texture analyzer, and image data information was acquired by industrial camera. The mechanical characteristics of the downward pressure work and the image features of the target shape change were extracted. The values of pressure work, the mechanical values and the image features were taken as inputs, and K-Nearest Neighbor classifier (KNN) was selected to construct the salt sea cucumber quality classification model. The classification accuracy rate was up to 77.22%, 88.8% and 88.8%, respectively. In order to improve the classification accuracy, the numerical features extracted above were normalized, which were taken as inputs for the quality classification model of salted sea cucumber based on KNN was established. The classification accuracy was 94.44%, which was proved the potential of this method for salted sea cucumber classification.