基于点云图的三维生猪模型重建技术

3D pig model reconstruction technique based on point-cloud image

目的:运用点云处理技术,重建生猪三维模型,为后期的生猪体尺自动测量和体质量估测研究提供必备条件,减少人工测量带来的疫病传播风险和应激反应,降低猪场经济损失。方法:开发多视角视频数据采集平台,实现2个深度相机分别从不同角度同步采集视频数据,经处理获取2个视角的点云数据,通过研究点云图的精准匹配技术,重建生猪三维模型。结果:通过对生猪点云图的精配准能够生成准确且完整的生猪三维模型,相较于粗配准的三维模型体尺测量结果提升了0.6%。结论:三维生猪模型为生猪后期的体尺测量提供一种高效、非接触式的解决方案,具有广阔的应用前景和潜在的实际价值。

Objective:Using point-cloud processing technology,the 3D model of live pigs was reconstructed,which provided necessary conditions for the research of automatic body size measurement and weight estimation of live pigs in the later stage,reduced the risk of disease transmission and stress reaction caused by manual measurement,and reduced the economic loss of pig farms.Methods:The multi-view video data acquisition platform was developed to realize the simultaneous acquisition of video data from different angles by two depth cameras,and the point-cloud data of two perspectives was obtained after processing.The 3D model of live pigs was reconstructed by studying the accurate matching technology of point-cloud images.Results:The accurate and complete 3D model of live pigs could be generated by the fine registration of the point-cloud image of live pigs,and the body size measurement results were improved by 0.6%compared with the coarse registration of the 3D model.Conclusion:3D pig model provided an efficient and non-contact solution for body size measurement in the later stage of live pigs,which had broad application prospects and potential practical value.