ChinaSpec: a network for long-term in situ measurements of solar-induced fluorescence and reflectance in China

Abstract

Extensive research has been conducted on assessing grasp stability, a crucial prerequisite for achieving optimal grasping strategies, including the minimum force grasping policy. However, existing works employ basic feature-level fusion techniques to combine visual and tactile modalities, resulting in the inadequate utilization of complementary information and the inability to model interactions between unimodal features. This work proposes an attention-guided cross-modality fusion architecture to comprehensively integrate visual and tactile features. This model mainly comprises convolutional neural networks (CNNs), self-attention, and cross-attention mechanisms. In addition, most existing methods collect datasets from real-world systems, which is time-consuming and high-cost, and the datasets collected are comparatively limited in size. This work establishes a robotic grasping system through physics simulation to collect a multimodal dataset. To address the sim-to-real transfer gap, we propose a migration strategy encompassing domain randomization and domain adaptation techniques. The experimental results demonstrate that the proposed fusion framework achieves markedly enhanced prediction performance (approximately 10%) compared to other baselines. Moreover, our findings suggest that the trained model can be reliably transferred to real robotic systems, indicating its potential to address real-world challenges.

Application prospects for combining active and passive observations of chlorophyll fluorescence

“…ChinaSpec (Zhang et al, 2020) (Frankenberg et al, 2011;Guanter et al, 2012Guanter et al, , 2014Porcar-Castell et al, 2014;Verma et al, 2017)。 然而, 目前这种经验关系缺乏机理解释 (Magney et al, 2019a), 并且SIF观测值和GPP之间的关系会随 着 时 空 尺 度 的 变 化 而 变 化 (Cheng et al, 2013;Zarco-Tejada et al, 2016;Goulas et al, 2017; (Cui et al, 2016;Miao et al, 2018), Wyber et al, 2017;Miao et al, 2018;Campbell et al, 2019;Magney et al, 2019a) 夫琅和费吸收暗线 (Fournier et al, 2012;Burkart et al, 2015;Yang et al, 2015), 这种检索波段的单一性 也使得SIF信号的解释存在不确定性。另外, 由于卫 星每天最多只能获得一个SIF观测值, 有效的SIF观 测值主要发生在晴天特定时间段 (Hu et al, 2018;Yang et al, 2018)。因此, 光照条件、植被结构、背 景反射和大气效应等都会影响荧光反演的结果 (Guanter et al, 2010;Zarco-Tejada et al, 2013)。为了 有效降低卫星SIF观测不确定性所带来的影响, 地 面荧光观测网络的建立、拓展以及荧光模型的构建 是必不可少的 (Zarco-Tejada et al, 2013;Atherton et al, 2016;Drusch et al, 2016;Grossmann et al, 2018;Raczka et al, 2019;Zhang et al, 2020)。 作用下与卫星SIF产品进行比较验证(van der Tol et al, 2014;Lee et al, 2015;Atherton et al, 2016;Raczka et al, 2019) 同的变化 (Huang et al, 2007;Pinto et al, 2016;Vilfan et al, 2016;Yang et al, 2018;Atherton et al, 2019)。 为了将SIF信号与植被生产力关联起来, 有必要对荧 光光谱进行重建, 建立全波段的SIF光谱曲线…”

“…et al, 2018; 章钊颖等, 2019), 但由于是基 于离散点的观测所以在空间上极不连续(Sun et al, 2018); GOME-2生产的SIF数据开始于2007年并持 续至今, 时间尺度较长因此适合研究长时间序列上 SIF值的变化, 但是空间分辨率比较低(40 km × 80 km, 其 与 卫 星 SIF 监 测 存 在 时 空 不 匹 配 问 题 (Zhang et al, 2020)。 自2010年近地面SIF自动观测系 统(Daumard et al, 2010)提出以来, 近地面SIF观测 已经被证明可以弥补卫星数据与EC通量塔数据之 间的差距(Porcar-Castell et al, 2015)。其观测数据与 TROPOMI、OCO-2等卫星SIF产品具有良好的相关性(Magney et al, 2019a;Zhang et al, 2020), …”