由于背景环境复杂,检测物体易受部分遮挡、天气以及光线变化等因素的影响,传统目标检测方法存在提取特征难、检测准确率低、检测耗时长等缺陷.为了改善传统目标检测方法存在的缺陷,实现快速准确的目标检测,提出了一种基于快速区域卷积...由于背景环境复杂,检测物体易受部分遮挡、天气以及光线变化等因素的影响,传统目标检测方法存在提取特征难、检测准确率低、检测耗时长等缺陷.为了改善传统目标检测方法存在的缺陷,实现快速准确的目标检测,提出了一种基于快速区域卷积神经网络(faster regions with convolutional neural network,Faster-RCNN)算法的轻量化改进方法,即针对算法Inception-V2特征提取网络进行轻量化改进,并以带泄露线性整流(leaky rectified linear unit,Leaky ReLU)作为激活函数,解决使用线性整流(rectified linear unit,ReLU)激活函数存在的神经元输入为负数时输出为0的问题.基于上述改进方法,选择沙滩废弃物的检测为案例以验证方法的有效性,并且结合不同特征提取网络在检测沙滩废弃物时的表现,对比了SSD(single shot multibox detector)与Faster-RCNN算法.实验结果表明:所提改进算法在实际检测中有较好的综合性能,且相比原算法Faster-RCNN_Inception-V2,轻量化改进后的Inception-V2特征提取网络卷积计算量减少51.8%,模型训练耗时缩短了9.1%,检测耗时减少了10.9%,各类别AP的平均值(mean average precision,mAP)增加了1.02%,可见所提的改进方法能够有效提高目标检测的准确率,减少检测耗时,并在沙滩废弃物检测上得到成功应用,为海滨城市的沙滩清理维护提供了技术支持与保障.展开更多
A reliable ultrasound-assisted extraction （UAE） method combined with HPLC-UV for quantification of eight active alkaloids in fruits of Macleaya cordata （Willd） R. Br. was developed. The optimization conditions of ...A reliable ultrasound-assisted extraction （UAE） method combined with HPLC-UV for quantification of eight active alkaloids in fruits of Macleaya cordata （Willd） R. Br. was developed. The optimization conditions of UAE were obtained by using Box-Behnken design of response surface methodology. Chromatography was carried out using a Kromasil C18 column by gradient elution with 0.1% phosphoric acid aqueous solution for HPLC-UV. All calibration curves showed good linear correlation coefficients （R^2〉0.999 6） and recoveries （from 97.3% to 104.9%） were acceptable. 1,1-diphenyl-2-picrylhydrazyl （DPPH） method was employed to test the antioxidant activity of the extract from the samples. The proposed method was successfully applied to quantifying eight components in nine samples of M.cordata, and significant variations of alkaloid contents and antioxidant aetivity of the samples from different habitats were demonstrated. It presents a powerful proof for the selection of the best sources to extract eight kinds of alkaloids.展开更多
文摘由于背景环境复杂,检测物体易受部分遮挡、天气以及光线变化等因素的影响,传统目标检测方法存在提取特征难、检测准确率低、检测耗时长等缺陷.为了改善传统目标检测方法存在的缺陷,实现快速准确的目标检测,提出了一种基于快速区域卷积神经网络(faster regions with convolutional neural network,Faster-RCNN)算法的轻量化改进方法,即针对算法Inception-V2特征提取网络进行轻量化改进,并以带泄露线性整流(leaky rectified linear unit,Leaky ReLU)作为激活函数,解决使用线性整流(rectified linear unit,ReLU)激活函数存在的神经元输入为负数时输出为0的问题.基于上述改进方法,选择沙滩废弃物的检测为案例以验证方法的有效性,并且结合不同特征提取网络在检测沙滩废弃物时的表现,对比了SSD(single shot multibox detector)与Faster-RCNN算法.实验结果表明:所提改进算法在实际检测中有较好的综合性能,且相比原算法Faster-RCNN_Inception-V2,轻量化改进后的Inception-V2特征提取网络卷积计算量减少51.8%,模型训练耗时缩短了9.1%,检测耗时减少了10.9%,各类别AP的平均值(mean average precision,mAP)增加了1.02%,可见所提的改进方法能够有效提高目标检测的准确率,减少检测耗时,并在沙滩废弃物检测上得到成功应用,为海滨城市的沙滩清理维护提供了技术支持与保障.
基金Project(20576142) supposed by the National Natural Science Foundation of China Project(2009DFA31270) supported by the International Cooperation Project of Ministry of Science and Technology of China
文摘A reliable ultrasound-assisted extraction （UAE） method combined with HPLC-UV for quantification of eight active alkaloids in fruits of Macleaya cordata （Willd） R. Br. was developed. The optimization conditions of UAE were obtained by using Box-Behnken design of response surface methodology. Chromatography was carried out using a Kromasil C18 column by gradient elution with 0.1% phosphoric acid aqueous solution for HPLC-UV. All calibration curves showed good linear correlation coefficients （R^2〉0.999 6） and recoveries （from 97.3% to 104.9%） were acceptable. 1,1-diphenyl-2-picrylhydrazyl （DPPH） method was employed to test the antioxidant activity of the extract from the samples. The proposed method was successfully applied to quantifying eight components in nine samples of M.cordata, and significant variations of alkaloid contents and antioxidant aetivity of the samples from different habitats were demonstrated. It presents a powerful proof for the selection of the best sources to extract eight kinds of alkaloids.
