基于Kinect相机的香梨重量预测方法  

Prediction method of fragrant pear weight based on Kinect camera

作　　者：张润芝 张晓[1] 吴刚[1] ZHANG Runzhi;ZHANG Xiao;WU Gang(College of Information Engineering,Tarim University,Alar,Xinjiang 843300,China)

机构地区：[1]塔里木大学信息工程学院,新疆阿拉尔843300

出　　处：《食品与机械》2023年第9期77-82,88,共7页Food and Machinery

基　　金：国家自然科学基金地区科学基金项目(编号:31960503);新疆生产建设兵团财政科技计划项目(编号:2021DB001)。

摘　　要：目的:快速获取香梨重量。方法:通过Kinect相机采集香梨的RGB-D图像并将其转化为点云数据;经点云预处理及点云插值后生成香梨模型;再利用微元法思想计算香梨模型的体尺参数,通过实验验证方法的准确性并修正误差;最后通过香梨体积预测香梨重量。结果:误差修正后体积的平均相对误差为2.8%;重量的平均相对误差为1.96%。结论:在距地面50 cm处采集图像的前提下,香梨各体尺参数测量值的平均相对误差均不超过5%,使用Kinect相机测量香梨体尺参数具有可行性。Objective:Aiming to obtain the weight of fragrant pear quickly to provide a basis for developing the fragrant pear grading device.Methods:This method acquired RGB-D images of fragrant pear by Kinect camera and converted them into point cloud data.The point cloud data was pre-processed and interpolated to generate a fragrant pear model.Then calculated the size parameter of the fragrant pear model.Finally,using the fragrant pear's volume predicted its weight.Results:Experimental results showed that the average relative error of the volume was 2.8%.Then the volume of fragrant pears was calculated by the error-compensated measurement method of the body scale parameter,and its weight was predicted and compared with the actual weight.The experimental results showed that the average relative error of the predicted weight was 1.96%.Conclusion:The fragrant pear quality prediction method provides a reliable reference for developing fragrant pear grading equipment.

关 键 词：RGB-D 三维点云 点云预处理 香梨建模 RANSAC 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] S661.2[自动化与计算机技术—计算机科学与技术]