摘要
为了给海地瓜识别工作提供数据参考,解决传统数据集构建中存在的数据重复率高的问题,笔者提出基于3D Studio Max的海地瓜数据集构建方法。该方法首先利用硬件设备采集海地瓜的三维数据,并对初始数据进行降噪和去重处理,然后构建海地瓜模型,最后以海地瓜模型为数据内容进行分类存储,实现海地瓜数据集的构建。实验结果表明:该方法的数据重复率降低了5.43%,能够有效提升识别的速度。
In order to provide data reference for the Haitian melon identification work and solve the problem of high data repetition rate in the traditional data set construction,the author proposes a Haitian melon dataset construction method based on 3D Studio Max.This method first uses hardware equipment to collect 3D data of Haitiangua,and performs noise reduction and deduplication processing on the initial data,then builds a Haitiangua model,and finally uses the Haitiangua model as the data content to classify and store the data to realize the construction of the Haitiangua dataset.The experimental results show that the data repetition rate of this method is reduced by 5.43%,which can effectively improve the recognition speed.
作者
王颖
WANG Ying(Xiamen Institute of Technology,Xiamen Fujian 361000,China)
出处
《信息与电脑》
2021年第5期44-46,共3页
Information & Computer
