在多目标跟踪任务中,外界噪声的干扰会导致传统方法的系统建模不可靠,从而降低目标位置预测的准确性;而密集人群引起的拥挤和遮挡问题则会严重影响目标外观的可靠性,导致错误的身份关联.为了解决这些问题,本文提出一种多目标跟踪算法Ecs...在多目标跟踪任务中,外界噪声的干扰会导致传统方法的系统建模不可靠,从而降低目标位置预测的准确性;而密集人群引起的拥挤和遮挡问题则会严重影响目标外观的可靠性,导致错误的身份关联.为了解决这些问题,本文提出一种多目标跟踪算法Ecsort.该算法在传统运动预测的基础上,引入噪声补偿模块,降低噪声干扰引起的误差,提高位置预测的准确性.其次,引入特征相似度匹配模块,通过学习目标的判别性外观特征,并结合运动线索和判别性外观特征的优势,从而实现精确的身份关联.通过在多目标跟踪基准数据集上进行的大量实验结果表明,与基线模型相比,该方法在MOT17测试集上的IDF1 (ID F1 score)、HOTA (higher order tracking accuracy)、AssA(association accuracy)、DetA (detection accuracy)分别提高了1.1%、0.5%、0.6%、0.3%,在MOT20测试集上的IDF1、HOTA、AssA、DetA分别提高了2.3%、1.9%、3.4%、0.2%.展开更多
[Objective] The aim of this study was to evaluate the effects of soil amend- ment composed of biochar and inorganic mineral material on growth and develop- ment, yield and output value of flue-cured tobacco in South A...[Objective] The aim of this study was to evaluate the effects of soil amend- ment composed of biochar and inorganic mineral material on growth and develop- ment, yield and output value of flue-cured tobacco in South Anhui Province. [Method] A field plot experiment was conducted. The agronomic traits, yield and output value, as well as appearance quality and flavor of flue-cured tobacco were evaluated. [Re- suit] The soil amendments composed of biochars and inorganic mineral materials could significantly increase plant height, stem diameter, leaf size, yield and output of flue-cured tobacco, as well as proportion of first-grade tobacco leaves at the middle and late growth period. The yield in the X3 treatment group （70% T20 ＋ 30% ZC） was highest, and it was higher than that in the control group by 398 kg/hm2. Com- pared with that in the control group, the output of flue-cured tobacco in the X3 treatment group was increased by 10 290 yuan/hm2. In terms of appearance quality and flavor, the flue-cured tobacco leaves in the soil amendment treatment groups were all better than those in the control group. [Conclusion] The application of soil amendment composed of biochar and inorganic mineral material is an effective mea- sure to improve tobacco-growing soil, promote tobacco growth and development, im- prove tobacco yield and output and improve tobacco leaf quality in South Anhui re- gion.展开更多
[Objective] The aim was to explore tobacco varieties which satisfy tobacco formula and is suitable to be grown in ecotopes simultaneously. [Method] Interior and exterior qualities of tobacco were compared and identifi...[Objective] The aim was to explore tobacco varieties which satisfy tobacco formula and is suitable to be grown in ecotopes simultaneously. [Method] Interior and exterior qualities of tobacco were compared and identified based on agronomic and economic characters of tobacco in fields. [Result] Yunshai No.1 was better in both agronomic and economic characters. Specifically, the fragrance was dominated by empyreumatique and aroma; the highly-qualified variety is fragrant and sweet, with little irritation, which gives the tobacco smoke a light, mild flavor. As for Virgini- a934, economic characters were better, but agronomic characters were just ordinary. Similar to Yunshai No.l, dominated by empyreumatique and aroma, the variety was with distinctive resin and herbaceous fragrance. What＇s more, it tasted good with high quality and usability. Cunsanpi was ordinary in agronomic characters, and poor in economic characters, with coordinated quality and chemical composition. It is fra- grant and the smoke is clean. However, the quality was not so high and it tasted not so good. Tengruan No.2 was good in agronomic characters, but poor in eco- nomic characters, which gives delicate fragrance, but not clean enough. Waiweiba and Mijie Qinggeng neither taste well nor smell sweet, performing ordinary in agro- nomic and economic characters. Hence, the two varieties seem bad in quality and ranks poorly. [Conclusion] In terms of sensory quality, the varieties from high to low were Yunshai No.1, Virginia 934, Cunsanpi, Tengruan No.2, Waiweiba and Mijie Qinggeng.展开更多
文摘在多目标跟踪任务中,外界噪声的干扰会导致传统方法的系统建模不可靠,从而降低目标位置预测的准确性;而密集人群引起的拥挤和遮挡问题则会严重影响目标外观的可靠性,导致错误的身份关联.为了解决这些问题,本文提出一种多目标跟踪算法Ecsort.该算法在传统运动预测的基础上,引入噪声补偿模块,降低噪声干扰引起的误差,提高位置预测的准确性.其次,引入特征相似度匹配模块,通过学习目标的判别性外观特征,并结合运动线索和判别性外观特征的优势,从而实现精确的身份关联.通过在多目标跟踪基准数据集上进行的大量实验结果表明,与基线模型相比,该方法在MOT17测试集上的IDF1 (ID F1 score)、HOTA (higher order tracking accuracy)、AssA(association accuracy)、DetA (detection accuracy)分别提高了1.1%、0.5%、0.6%、0.3%,在MOT20测试集上的IDF1、HOTA、AssA、DetA分别提高了2.3%、1.9%、3.4%、0.2%.
基金Supported by Special Project for Shanghai and Anhui Modern Tobacco Agricultural High-tech Demonstration Park(CF56.1-ZJ1)~~
文摘[Objective] The aim of this study was to evaluate the effects of soil amend- ment composed of biochar and inorganic mineral material on growth and develop- ment, yield and output value of flue-cured tobacco in South Anhui Province. [Method] A field plot experiment was conducted. The agronomic traits, yield and output value, as well as appearance quality and flavor of flue-cured tobacco were evaluated. [Re- suit] The soil amendments composed of biochars and inorganic mineral materials could significantly increase plant height, stem diameter, leaf size, yield and output of flue-cured tobacco, as well as proportion of first-grade tobacco leaves at the middle and late growth period. The yield in the X3 treatment group （70% T20 ＋ 30% ZC） was highest, and it was higher than that in the control group by 398 kg/hm2. Com- pared with that in the control group, the output of flue-cured tobacco in the X3 treatment group was increased by 10 290 yuan/hm2. In terms of appearance quality and flavor, the flue-cured tobacco leaves in the soil amendment treatment groups were all better than those in the control group. [Conclusion] The application of soil amendment composed of biochar and inorganic mineral material is an effective mea- sure to improve tobacco-growing soil, promote tobacco growth and development, im- prove tobacco yield and output and improve tobacco leaf quality in South Anhui re- gion.
基金Supported by China Tobacco Yunnan Industrial Co.,Ltd(2010YL02)~~
文摘[Objective] The aim was to explore tobacco varieties which satisfy tobacco formula and is suitable to be grown in ecotopes simultaneously. [Method] Interior and exterior qualities of tobacco were compared and identified based on agronomic and economic characters of tobacco in fields. [Result] Yunshai No.1 was better in both agronomic and economic characters. Specifically, the fragrance was dominated by empyreumatique and aroma; the highly-qualified variety is fragrant and sweet, with little irritation, which gives the tobacco smoke a light, mild flavor. As for Virgini- a934, economic characters were better, but agronomic characters were just ordinary. Similar to Yunshai No.l, dominated by empyreumatique and aroma, the variety was with distinctive resin and herbaceous fragrance. What＇s more, it tasted good with high quality and usability. Cunsanpi was ordinary in agronomic characters, and poor in economic characters, with coordinated quality and chemical composition. It is fra- grant and the smoke is clean. However, the quality was not so high and it tasted not so good. Tengruan No.2 was good in agronomic characters, but poor in eco- nomic characters, which gives delicate fragrance, but not clean enough. Waiweiba and Mijie Qinggeng neither taste well nor smell sweet, performing ordinary in agro- nomic and economic characters. Hence, the two varieties seem bad in quality and ranks poorly. [Conclusion] In terms of sensory quality, the varieties from high to low were Yunshai No.1, Virginia 934, Cunsanpi, Tengruan No.2, Waiweiba and Mijie Qinggeng.
