Search

Neural Radiance Fields (NeRF) has shown great success in novel view synthesis due to its state-of-the-art quality and flexibility. However, NeRF requires dense input views (tens to hundreds) and a long training time (hours to days) for a single scene to generate high-fidelity images. Although using the voxel grids to represent the radiance field can significantly accelerate the optimization process, we observe that for sparse inputs, the voxel grids are more prone to overfitting to the training views and will have holes and floaters, which leads to artifacts. In this paper, we propose VGOS, an approach for fast (3-5 minutes) radiance field reconstruction from sparse inputs (3-10 views) to address these issues. To improve the performance of voxel-based radiance field in sparse input scenarios, we propose two methods: (a) We introduce an incremental voxel training strategy, which prevents overfitting by suppressing the optimization of peripheral voxels in the early stage of reconstruction. (b) We use several regularization techniques to smooth the voxels, which avoids degenerate solutions. Experiments demonstrate that VGOS achieves state-of-the-art performance for sparse inputs with super-fast convergence. Code will be available at https://github.com/SJoJoK/VGOS., Comment: IJCAI 2023 Accepted (Main Track)

Mucky soft soil with weak properties was usually improved by the chemical stabilization owing to the high cost of the situ replacement, and the main chemical binder used was Portland cement (PC). However, with the needs of environmental protection, energy conservation and emissions reduction, the research of low-carbon and high-efficiency soft soil treatment technology will be the development trend of modern engineering construction. The combined binder of reactive magnesia (MgO) and carbon dioxide (CO2) was regarded as a new curing agent instead of traditional PC, and the mass carbonation technology as an innovative idea was put forward to reinforce soft soil. In this view, the field mass stabilization was carried out and the single-point test was analyzed. Research results show that the temperature from the carbonation reaction is higher than that from the hydration reaction, and the temperature, strength and carbonation degree of the upper foundation soils increase significantly compared with the lower foundation soils, while the water content and pH value of the upper foundation soils are less than those of the lower foundation soils; the smaller the distance from the CO2 gas supply is, the higher the carbonation temperature, resilience modulus, penetration resistance, strength and carbonation degree are; the pipe-cover mass carbonation–stabilization technology can realize the carbonation and reinforcement of the on-site weak soils, but it should be noted that the CO2 should be injected at the bottom. The maximum ventilation radius should be 60 cm and the distance between the CO2 pipes should be less than 1.2 m in practical application. The bearing capacities of all regions are above 200 kPa, and the resilient moduli are about 20 MPa. The experimental research of mass stabilization based on MgO-CO2 carbonation technology has groundbreaking scientific significance and application prospects in the emission reduction and sustainable development of civil engineering.

Using the technique of laser flash photolysis, it has been demonstrated that the interaction of SO − 4 • with methionine in aqueous solution at pH 1.0–7.0 results in the formation of the H abstraction radicals Met(-H) · and the dimer of three-electron-bonded radical cations Met − 2 • [S] with maximum absorption at 300 and 480 nm respectively. The characteristic absorption at 480 nm completely disappeared at pH >7.0. The mechanism has been deduced and the kinetic parameters have been determined.

The first year operation of HL-1, a medium size, circular tokamak ( R =102 cm, a = 20 cm) was highlighted by the achievement of plasma current duration in excess of 600 ms ( I p = 100 kA). This is partially attributed to the low Z eff discharges made possible by wall conditioning using Taylor discharge cleaning in hydrogen (TDC), accompanied by baking. Approximately 30 h of TDC and PDC were required to obtain this low Z eff (- 2) tokamak discharges. During the initial clean-up of the wall, a great deal ( ∼ 50 monolayers) of O and C was removed. In this paper, we present: (1) a description of the TDC system and optimization of TDC operation, (2) some behaviour of the partial pressure of water vapour as function of filling pressure, plasma current, electron density (temperature), and discussion of the mechanism for removing oxygen with a simple model called the “three step model” to remove oxygen, compared with the experimental results. The sticking probability, P st is estimated to be around 0.3 for GH-39 steel bellows for clean conditions in HL-1.