摘要
为构建绣球菌子实体工厂化生长模型,揭示绣球菌形态和产量的形成过程,并实现绣球菌子实体三维可视化及表型参数的自动提取,以闽绣1号为材料,开展2次工厂化栽培试验.每隔1 d,手动采集绣球菌子实体农艺指标,并在成熟期获取产量和子实体多视角图像数据.通过SPSS软件分析,建立绣球菌子实体高度、长度、宽度、产量及整个菌包质量等农艺指标随生长天数的模型,模型决定系数范围为0.935～0.995,经检验模型预测效果较好,平均相对误差范围为2.39％～8.09％.然后,以绣球菌子实体的高度、长度、宽度为自变量无损地评估绣球菌生产过程的产量动态变化.最后,基于绣球菌子实体多视角图片数据,实现了绣球菌子实体三维可视化以及子实体高度、长度、宽度、表面积、体积等表型参数的自动提取,三维模型算法计算值与人工测量值或者计算值相比,平均相对误差均小于14％,为绣球菌高通量表型参数自动获取及优良新品种选育提供重要支撑.
Abstract
In order to establish a factory growth model for fruiting bodies of Sparassis crispa,reveal the formation pro-cess of the morphology and yield of fruiting bodies of S.crispa,realize three-dimensional visualization and automatic ex-traction of phenotypic parameters.Minxiu No.1 was used as experimental material,two factory cultivation experiments were carried out.The agronomic indexes of fruiting bodies of S.crispa were collected every 2 days,yield and multi-view image data were obtained at the maturity stage.Through SPSS software analysis,the model of the agricultural indicators such as the height,length,width,yield of fruiting bodies and mass of the whole S.crispa package with the growth days were established,the determination coefficients of the simulation model were between 0.935 and 0.995,the prediction ef-fects were good,the mean relative errors ranged from 2.39％to 8.09％.Then,the yield of fruiting bodies of S.crispa were evaluated non-destructively by using the fruiting bodies height,length and width as independent variables.Finally,based on multi-view picture of fruiting bodies of S.crispa,three-dimensional visualization and the automatic extraction of the height,length,width,surface area,volume of fruiting bodies were realized,compared with manual measurements or cal-culated value,the mean relative errors were less than 14％,which provides important support for high-throughput pheno-typic parameters obtained automatically and excellent new variety breeding of S.crispa.
维普
万方数据