基于手机拍摄图像猪肉新鲜度的检测

Research on Detection of Pork Freshness System Based on Mobile Photograph Image

首先对手机拍摄的猪肉图像进行预处理,利用Otsu算法较好地分割出图像中的猪肉部分和背景,然后用颜色因子∣R-G∣+∣R-B∣的值对图像进行彩色分层从而计算出颜色区域比值,建立了用猪肉的色泽、黏度、弹性、氨气、硫化氢、表面菌落总数、颜色区域比等7个特征为输入的BP网络分类模型。试验结果表明,该方法能够较好地对猪肉新鲜度进行检测。

Pork images obtained were pre-processed firstly by Otsu algorithm, pork samples and background were segmented and isolated. Subsequently, the value of color factor IR-G I ＋ I R-BI which calculated the color region ratio was used to color layering the images. Finally, the pork freshness model was constructed based on BP neural network which inputted by the color, viscosity, elasticity, ammonia, hydrogen sulfide, the total number of surface colonies, color region ratio. Recognition test results showed that the method can detect pork freshness effectively.