首页|基于轮廓整形单元的原料肉3D可视化成像系统研究

基于轮廓整形单元的原料肉3D可视化成像系统研究

扫码查看
针对3D激光扫描实现不规则原料肉轮廓成像时存在扫描轮廓不完整、部分信息缺失、体积估测精度低等问题,根据原料肉物性学特性,本文设计了一种基于轮廓整形单元的3D可视化成像系统,实现不规则原料肉的形态调整,以优化不规则原料肉的成像性能.阐述了原料肉轮廓整形单元的工作原理,设计了样品驱动传输单元、扫描外触发控制单元、可实现轮廓整形的成像检测平台等硬件模块,确定了整形单元中铰链螺栓旋转方向、电机旋转圈数和原料肉轮廓目标整形角度之间的关系.基于Halcon平台和C#语言编写了轮廓整形优化原料肉成像性能的3D可视化软件,利用点云处理模型重建算法和灰度膨胀(Gray dilation)孔洞补偿算法,实现不规则原料肉轮廓整形前后信息采集、数据分析及体积估测准确度对比.最后选用120块猪肉(后腿、里脊)在冷鲜和麻冻状态下验证整形单元对原料肉轮廓成像的优化性能.结果表明,与传输方向呈90°、180°、270°、360°时肉块扫描后成像精度均不小于90%,变异系数不大于3%.冷鲜肉和麻冻肉最佳整形角度范围分别为30°~50°和40°~60°,体积估测准确率从90%分别提高至94%和97%以上.整形后其轮廓形态可维持时间均大于6s,孔洞高度最大压缩比均小于0.77.研究证明,不规则原料肉经过轮廓整形单元后可有效提高成像性能,为后续基于轮廓成像实现不规则原料肉定量分切技术研发提供支持.
Design and Experimental Verification of 3D Visual Imaging System Based on Contour Shaping Unit
Aiming at 3D laser scanning to achieve irregular raw meat contour imaging,there are problems such as incomplete scanning contours,missing data,and low volume estimation accuracy.In light of these limitations,a 3D visual imaging system was presented based on the contour shaping unit.This system was designed to address the morphological characteristics of irregular raw meat,with the aim of optimizing the imaging performance of irregular raw meat.The operational methodology of the contour shaping apparatus was delineated,and the essential hardware modules,including the sample driving and transmission unit,scanning external trigger control unit,and imaging detection platform.Additionally,the relationship between the rotational orientation of the hinge bolt in the shaping apparatus,the number of motor rotations,and the desired contour angle of the raw material meat was determined.A 3D visualization software was ultimately developed on the Halcon platform by utilizing the C# language.The point cloud processing model reconstruction algorithm and gray dilation hole compensation algorithm were employed to facilitate the acquisition of information,analysis of data,and comparison of volume estimation accuracy before and after contour shaping of irregular raw meat.This was done in order to validate contour shaping to optimize the imaging performance of meat.A total of 120 pieces of chilled and frozen pork(hind shank and loin)was employed to substantiate the enhanced functionality of the shaping unit for the imaging of raw meat contours.The results demonstrated that the post-scanning imaging accuracies of the meat pieces at 90°,180°,270° and 360° relative to the transmission direction were greater than 90%,and the coefficients of variation were no more than 3% .The optimal angle for shaping ranged from 30° to 50° for chilled meat and from 40° to 60° for frozen meat.The accuracy of volume estimation was improved from 90% to over 94%,and 97%,respectively.Following the shaping process,the contour of chilled and frozen meat morphology can be maintained for over 6 s,with a maximum compression ratio of hole height below 0.77.The research result demonstrated that the imaging performance of irregular raw meat can be significantly enhanced through the application of a contour shaping unit.This finding provided a valuable foundation for subsequent research and development efforts aimed at advancing quantitative slitting technology based on contour imaging for irregular raw meat.

meat recognition3D visualizationcontour shapinglaser scanningpoint cloud processingvolume estimation

卜令平、高国伟、乔真、田惠鑫、胡敬芳、张春晖、胡小佳、艾鑫、李侠、魏文松

展开 >

北京信息科技大学北京市传感器重点实验室,北京 100101

中国农业科学院农产品加工研究所农业农村部农产品加工综合性重点实验室,北京 100193

农业农村部农产品产地处理装备重点实验室,杭州 310058

淄博数字农业农村研究院,淄博 255035

北京信息科技大学现代测控技术教育部重点实验室,北京 100192

展开 >

肉品识别 3D可视化 轮廓整形 激光扫描 点云处理 体积估测

2024

农业机械学报
中国农业机械学会 中国农业机械化科学研究院

农业机械学报

CSTPCD北大核心
影响因子:1.904
ISSN:1000-1298
年,卷(期):2024.55(z1)