RESUMEN
We study multi-sensor fusion for 3D semantic segmentation that is important to scene understanding for many applications, such as autonomous driving and robotics. For example, for autonomous cars equipped with RGB cameras and LiDAR, it is crucial to fuse complementary information from different sensors for robust and accurate segmentation. Existing fusion-based methods, however, may not achieve promising performance due to the vast difference between the two modalities. In this work, we investigate a collaborative fusion scheme called perception-aware multi-sensor fusion (PMF) to effectively exploit perceptual information from two modalities, namely, appearance information from RGB images and spatio-depth information from point clouds. To this end, we first project point clouds to the camera coordinate using perspective projection. In this way, we can process both inputs from LiDAR and cameras in 2D space while preventing the information loss of RGB images. Then, we propose a two-stream network that consists of a LiDAR stream and a camera stream to extract features from the two modalities, separately. The extracted features are fused by effective residual-based fusion modules. Moreover, we introduce additional perception-aware losses to measure the perceptual difference between the two modalities. Last, we propose an improved version of PMF, i.e., EPMF, which is more efficient and effective by optimizing data pre-processing and network architecture under perspective projection. Specifically, we propose cross-modal alignment and cropping to obtain tight inputs and reduce unnecessary computational costs. We then explore more efficient contextual modules under perspective projection and fuse the LiDAR features into the camera stream to boost the performance of the two-stream network. Extensive experiments on benchmark data sets show the superiority of our method. For example, on nuScenes test set, our EPMF outperforms the state-of-the-art method, i.e., RangeFormer, by 0.9% in mIoU. Compared to PMF, EPMF also achieves 2.06× acceleration with 2.0% improvement in mIoU. Our source code is available at https://github.com/ICEORY/PMF.
RESUMEN
A luminescent 1D coordination polymer (CP) [Zn2L2(H2O)4]·H2O (1, H2L = 1-(4-carboxyphenyl)-1H-pyrazole-3-carboxylic acid) was prepared by a solvothermal method. 1 shows excellent fluorescence properties and has an obvious fluorescence "turn-on" phenomenon for saccharin (SAC), 2-thiazolidinethione-4-carboxylic acid (TTCA), and periodate (IO4-). Between 0 and 60 µM concentration range of SAC, the fluorescence enhancement efficiency (KEC) of 1 reaches 1.00 × 105 M-1 with the limit of detection (LOD) of 90 nM. 1 is the first CP-based sensing material for SAC detection. For TTCA detection, the KEC is 2.73 × 105 M-1 at the 25-80 µM concentration range, and the LOD is 33 nM, the lowest LOD among the sensors that detect TTCA at present. For IO4- ion detection, when the IO4- ion concentration ranges from 0 to 10 µM, the KEC is 2.34 × 105 M-1 and the LOD is as low as 39 nM. In order to better understand the sensing phenomenon, we also discuss in detail the sensing mechanisms for SAC, TTCA, and IO4- ions.
RESUMEN
The adsorption energy of oxygen-containing intermediates for the oxygen evolution reaction (OER) electrocatalysts plays a key role on their electrocatalytic performances. Rational optimization and regulation of the binding energy of intermediates can effectively improve the catalytic activities. Herein, the binding strength of Co phosphate to *OH was weakened by generating lattice tensile strain via Mn replacement, which modulated the electronic structure and optimized the reactive intermediates adsorption with active sites. The tensile-strained lattice structure and stretched interatomic distance were confirmed by X-ray diffraction and extended X-ray absorption fine structure (EXAFS) spectra measurements. The as-obtained Mn-doped Co phosphate exhibits excellent OER activity with an overpotential of 335 mV at 10 mA cm-2, which is much higher than pristine Co phosphate. In-situ Raman spectra and methanol oxidation reaction experiments demonstrated that Mn-doped Co phosphate with lattice tensile strain shows optimized *OH adsorption strength, and is favorable to structure reconstruction and form highly active Co oxyhydroxide intermediate during OER process. Our work provides insight into the effects of the lattice strain on the OER activity from the standpoint of intermediate adsorption and structure transformation.
