基于改进二进制粒子群算法的苹果多特征提取分级研究

为了提升对苹果分级的准确性,采用改进二进制粒子群算法对苹果多特征进行提取分级研究。首先建立苹果多特征提取量,包括大小、颜色、缺陷、形状特征;然后基于辅助搜索空间的二进制粒子群更新,对粒子位置增加状态翻转因子,根据收敛情况动态地获得单向翻转角度;接着通过Sigmoid函数、高斯函数对苹果多特征进行分级建模,确定了分段函数的参数值;最后给出了苹果分级的算法流程。实验仿真显示,该算法对苹果多特征提取分级的结果较其他算法更准确,且运行时间较少。

Multi-feature extraction of apple grading based on improved binary particle swarm optimization algorithm

In order to enhance the apple grade judgment, improved binary particle swarm optimization algorithm was proposed. Firstly, apple multi-feature extraction was established, which included size, color, defect and shape feature. Secondly, binary particle swarm space was updated with auxiliary search space, and position of particle was increased with state turnover factor, so that undirectional flip dynamic angle was made. Thirdly, model of the multi-feature classification of apple was established based on Sigmoid function and Gauss function, and the parameter values of the piecewise function was determined. Finally, the process was given. Simulation showed that improved binary particle swarm optimization was more accurate and had less time than other algorithm.