Swin-MFA: A Multi-Modal Fusion Attention Network Based on Swin-Transformer for Low-Light Image Human Segmentation.

In recent years, image segmentation based on deep learning has been widely used in medical imaging, automatic driving, monitoring and security. In the fields of monitoring and security, the specific location of a person is detected by image segmentation, and it is segmented from the background to analyze the specific actions of the person. However, in low-illumination conditions, it is a great challenge to the traditional image-segmentation algorithms. Unfortunately, a scene with low light or even no light at night is often encountered in monitoring and security. Given this background, this paper proposes a multi-modal fusion network based on the encoder and decoder structure. The encoder, which contains a two-branch swin-transformer backbone instead of the traditional convolutional neural network, fuses the RGB and depth features with a multiscale fusion attention block. The decoder is also made up of the swin-transformer backbone and is finally connected via the encoder with several residual connections, which are proven to be beneficial in improving the accuracy of the network. Furthermore, this paper first proposes the low light-human segmentation (LLHS) dataset of portrait segmentation, with aligned depth and RGB images with fine annotation under low illuminance, by combining the traditional monocular camera and a depth camera with active structured light. The network is also tested in different levels of illumination. Experimental results show that the proposed network has good robustness in the scene of human segmentation in a low-light environment with varying illumination. The mean Intersection over Union (mIoU), which is often used to evaluate the performance of image segmentation model, of the Swin-MFA in the LLHS dataset is 81.0, is better than those of ACNet, 3DGNN, ESANet, RedNet and RFNet at the same level of depth in a mixed multi-modal network and is far ahead of the segmentation algorithm that only uses RGB features, so it has important practical significance.

[Net carbon exchange and its driving factors of Haloxylon ammodendron plantation in the oasis-desert ecotone of Minqin, China]. / 民勤绿洲-è’æ¼ è¿‡æ¸¡å¸¦æ¢­æ¢­äººå·¥æž—å‡€ç¢³äº¤æ¢åŠå ¶å½±å“å› å­.

Carbon fluxes in a Haloxylon ammodendron plantation in the oasis-desert ecotone of Minqin was measured using an eddy covariance system. To provide scientific data for carbon source/sink assessment, we quantitatively analyzed the characteristics of CO2 flux and its driving factors in the growing season from May to October, 2018. The results showed that the trend of daily net carbon exchange in the growing season followed a symmetrical "U" shape curve. As to seasonality, bimodal curve was obvious. The plantation ecosystem was a carbon sink every month. The total carbon sequestrated was 34.38 g C·m-2, with the peak of 12.31 g C·m-2 in September and the lowest value of 0.89 g C·m-2 in July. The net carbon exchange in this ecosystem increased during the daytime with the increasing photosynthetically active radiation, consistent with the Michaelis-Menten rectangular hyperbola change. When the vapor pressure deficit was greater than 2.5 kPa, the increasing trend tended to flat. Ecosystem respiration increased exponentially with temperature, with temperature sensitivity being 1.7. Net carbon exchange in either day or night was significantly correlated with soil temperature through the whole growing season.

Factors influencing the natural regeneration of the pioneering shrub Calligonum mongolicum in sand dune stabilization plantations in arid deserts of northwest China.

Calligonum mongolicum is a successful pioneer shrub to combat desertification, which is widely used for vegetation restoration in the desert regions of northwest China. In order to reveal the limitations to natural regeneration of C. mongolicum by asexual and sexual reproduction, following the process of sand dune stabilization, we assessed clonal shoots, seedling emergence, soil seed bank density, and soil physical characteristics in mobile and stabilized sand dunes. Controlled field and pot experiments were also conducted to assess germination and seedling emergence in different dune soil types and seed burial depths. The population density of mature C. mongolicum was significantly different after sand dune stabilization. Juvenile density of C. mongolicm was much lower in stabilized sand dunes than mobile sand dune. There was no significant difference in soil seed bank density at three soil depths between mobile and stabilized sand dunes, while the emergence of seedlings in stabilized dunes was much lower than emergence in mobile dunes. There was no clonal propagation found in stabilized dunes, and very few C. mongolicum seedlings were established on stabilized sand dunes. Soil clay and silt content, air-filled porosity, and soil surface compaction were significantly changed from mobile sand dune to stabilized dunes. Seedling emergence of C. mongolicm was highly dependent on soil physical condition. These results indicated that changes in soil physical condition limited clonal propagation and seedling emergence of C. mongolicum in stabilized sand dunes. Seed bank density was not a limiting factor; however, poor seedling establishment limited C. mongolicum's further natural regeneration in stabilized sand dunes. Therefore, clonal propagation may be the most important mode for population expansion in mobile sand dunes. As a pioneer species C. mongolicum is well adapted to propagate in mobile sand dune conditions, it appears unlikely to survive naturally in stabilized sand dune plantations.