针对建筑垃圾物料的种类多、形貌易混淆等问题,构建了一种基于局部约束的视觉词袋(local constraint-bag of visual words,LC-BoVW)模型的建筑垃圾物料识别算法。首先,对建筑垃圾物料图像分块,分别提取局部颜色特征和局部二值模式特征;...针对建筑垃圾物料的种类多、形貌易混淆等问题,构建了一种基于局部约束的视觉词袋(local constraint-bag of visual words,LC-BoVW)模型的建筑垃圾物料识别算法。首先,对建筑垃圾物料图像分块,分别提取局部颜色特征和局部二值模式特征;考虑到图像分块特征的局部相似特性,构建LC-BoVW模型分别对目标图像的显著特征进行统计。然后,基于信息融合思想对特征统计量进行融合,形成图像的判别性特征并输入到分类器中进行物料的精确识别。最后,利用自建的5类建筑垃圾物料图像数据集进行实验,实验结果表明,所提算法能够快速有效地实现建筑垃圾物料识别,平均识别准确率可达到97.92%。展开更多
In recent years,karst construction projects in the built-up area of Wuhan(capital of Hubei Province,China)are increasing,and the karst geological disasters have aroused social concerns.The actual engineering projects ...In recent years,karst construction projects in the built-up area of Wuhan(capital of Hubei Province,China)are increasing,and the karst geological disasters have aroused social concerns.The actual engineering projects usually use shallow geophysical exploration methods to explore karst.This paper uses Spatial Auto-Correlation Method(SPAC)and electromagnetic Computerized Tomography(CT)to detect karst in urban built-up areas.Depending on the different physical properties of rock and soil,the SPAC method can better reveal the interface between soil and rock strata and the interface between soil layers.The electromagnetic CT method can identify strata according to the apparent absorption coefficient,which can better reveal the interface between soil and rock,the interface between the more intact and weathered rock.The SPAC method is mainly qualitative to measure the low-speed area,namely,the wrong geological body i.e.,karst cave,but also can detect the fracture zone or filling mode of karst cave,and at the same time,cannot use exploration holes or logging observation.The electromagnetic CT method can accurately detect the location and scale of the karst caves and has a higher accuracy detecting karst bands.In addition,exploration holes or well logging observations are also expected to be conducted,and their detection effect is greatly affected by lithology.展开更多
文摘针对建筑垃圾物料的种类多、形貌易混淆等问题,构建了一种基于局部约束的视觉词袋(local constraint-bag of visual words,LC-BoVW)模型的建筑垃圾物料识别算法。首先,对建筑垃圾物料图像分块,分别提取局部颜色特征和局部二值模式特征;考虑到图像分块特征的局部相似特性,构建LC-BoVW模型分别对目标图像的显著特征进行统计。然后,基于信息融合思想对特征统计量进行融合,形成图像的判别性特征并输入到分类器中进行物料的精确识别。最后,利用自建的5类建筑垃圾物料图像数据集进行实验,实验结果表明,所提算法能够快速有效地实现建筑垃圾物料识别,平均识别准确率可达到97.92%。
文摘In recent years,karst construction projects in the built-up area of Wuhan(capital of Hubei Province,China)are increasing,and the karst geological disasters have aroused social concerns.The actual engineering projects usually use shallow geophysical exploration methods to explore karst.This paper uses Spatial Auto-Correlation Method(SPAC)and electromagnetic Computerized Tomography(CT)to detect karst in urban built-up areas.Depending on the different physical properties of rock and soil,the SPAC method can better reveal the interface between soil and rock strata and the interface between soil layers.The electromagnetic CT method can identify strata according to the apparent absorption coefficient,which can better reveal the interface between soil and rock,the interface between the more intact and weathered rock.The SPAC method is mainly qualitative to measure the low-speed area,namely,the wrong geological body i.e.,karst cave,but also can detect the fracture zone or filling mode of karst cave,and at the same time,cannot use exploration holes or logging observation.The electromagnetic CT method can accurately detect the location and scale of the karst caves and has a higher accuracy detecting karst bands.In addition,exploration holes or well logging observations are also expected to be conducted,and their detection effect is greatly affected by lithology.
