基于自适应寻优遗传算法的牛肉变质区域识别方法

Recognition of Beef Deterioration Areas Based on Adaptive Optimal Genetic Algorithms

摘要:

针对传统方法对牛肉变质区域识别的准确率低、用时长的问题，提出基于自适应寻优遗传算法的牛肉变质区域识别方法。将牛肉检测光谱数据子集作为染色体，引用二进制编码，计算群个体编码及种群初始化；通过适应度函数优化交叉与变异操作，为种群的不同个体计算单独的交叉概率与变异概，输出最终牛肉变质区域的检测光谱数据。实验数据表明，所提方法仅迭代40次可完成牛肉变质区域识别，且平均识别准确率为95.8%，识别用时为1.7 s，与两种传统方法相比，识别精度分别提高了28.72%和20.34%，识别耗时分别缩短了2.09 s和4.13 s；由此得出结论，所提方法具有收敛速度快，且适应度均值较高，全局搜索能力强；所提方法在识别牛肉变质区域方面具有识别率高、用时短的优势，具有较高的可靠性、科学性和可行性。

Abstract:

To solve the problem of low accuracy and long time of traditional methods for beef deterioration area identification, a method based on adaptive optimization genetic algorithm was investigated. The subset of beef detection spectrum data was used as chromosome, and binary coding was used to calculate individual coding and population initialization. The fitness function was used to optimize crossover and mutation operation, and the individual crossover probability and mutation probability were calculated for different individuals of the population. The final detection spectrum data of beef deterioration area was output. The resultas showed that 40 iterations could conduct beef deterioration area recognition the proposed method. The average recognition accuracy was 95.8%, and the recognition time was 1.7 s. Compared to two traditional methods, the recognition accuracy in this work was improved by 28.72% and 20.34%, respectively, and the recognition time was shortened by 2.09 s and 4.13 s, respectively. The proposed method had the advantages of high recognition rate, short time, high reliability, scientificity and feasibility in identifying beef spoilage areas.