In the detection of beef physiological maturity, correct segmentation of the cartilage region at the end of the spine and selection of the cartilage region characteristic parameters are two important steps that will directly affect the accuracy of detection. In this study, the preprocessed spine image collected being converted into two binary images, geometric invariant moments and Hopfield are used to complete the automatic segmentation and recognition of cartilage region, with 、 and as beef physiological maturity evaluation standard. The average accuracy rate is 88% in detecting 20 samples of physiological maturity levels at A, B, C, D and E respectively. The result shows that the method can effectively eliminate the effect of bone chip and fat granule in the process of beef cutting, and has improved beef detection.%在牛肉生理成熟度的检测中,脊椎骨末端软骨区域的正确分割和软骨区域特征参数的选取是非常重要的2个步骤,将直接影响检测的准确度.本研究首先对采集到的脊骨图像进行预处理将其转换为二值图像,然后利用几何不变矩及Hopfield神经网络完成对软骨区域的自动分割和识别,并选取内角方差、凹凸度和区域密集性作为牛肉生理成熟度的评判标准.在对生理成熟度分别为A、B、C、D、E级的各20具牛肉样本进行实验检测的过程中,平均准确率达到了92%.结果表明:该方法可以有效的消除牛肉切割过程中产生的骨屑及脂肪粒对于检测结果的影响,具有较好的检测效果.
展开▼
