BZBeibei Zhou

Beibei Zhou

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Beibei Zhou has a total of 42 co-authors. Here are the top co-authors by number of shared publications:

  • Lijun Su

    State Key Laboratory of Eco-Hydraulics in Northwest Arid Region, Xi’an University of Technology, Xi’an 710048, China

    3 shared publications
  • Manli Duan

    Institute of Water Resources and Hydroelectric Engineering Xi'an Univ. of Technology Xi'an Shaanxi 710048 China

    3 shared publications
  • Quanjiu Wang

    State Key Laboratory of Eco‐hydraulic in Northwest Arid Region Xi'an Univ. of Technology Xi'an Shaanxi 710048 China

    3 shared publications
  • Dapeng Ye

    College of Mechanical and Electrical Engineering, Fujian Agriculture and Forestry University

    2 shared publications
  • Yuhua Zhang

    State Key Laboratory of Eco-hydraulics in Northwest Arid Region of China, Xi'an University of Technology, Xi'an 710048, China

    2 shared publications
  • Yuyang Shan

    State Key Laboratory of Eco‐hydraulic in Northwest Arid Region Xi'an Univ. of Technology Xi'an Shaanxi 710048 China

    2 shared publications

Latest Publications

  • Open Access

    Research on Rice Pest Detection and Classification Based on YOLOv5 and Transformer Combination

    by Qiaonan Yang, Yayong Chen, Qing Hai, Sehar Razzaq, Yiming Cui, Xingwang Wang, Beibei Zhou

    The significant differences in insects trapped by pest detection lamps lead to low classification accuracy of existing models for rice pests. To address this issue, this paper proposes a small pest target detection and classification model (ViT-YOLOv5p) by integrating the YOLO backbone and Transformer module. First, the number of training samples is expanded through data augmentation during model training. Furthermore, appropriate noise data are introduced to enhance the robustness and generalization ability of the model. Before detection and classification, image cutting and stitching strategies are adopted to improve the detection accuracy of small objects. The bounding box of the pest is determined by the YOLO backbone, and the corresponding region is fed into the Transformer model to obtain the classification result. Finally, YOLOv5, Faster R-CNN, YOLOv4, and the proposed ViT-YOLOv5p are trained on the same dataset, with average detection time (ADT) and classification accuracy employed as evaluative metrics. The results show that ViT-YOLOv5p achieves the highest classification accuracy of 91.89% with an ADT of 50.41 ms. Compared with the commonly used Faster R-CNN, YOLOv5, and YOLOv4 models, the accuracy is improved by 1.50%, 8.71%, and 9.74%, respectively. This study provides a reference for agricultural pest detection, automatic insect classification systems, and deep learning-based detection of small agricultural targets.

  • Open Access

    A method of deep network auto-training based on the MTPI auto-transfer learning and a reinforcement learning algorithm for vegetation detection in a dry thermal valley environment

    by Yayong Chen, Beibei Zhou, Chen Xiaopeng, Changkun Ma, Lei Cui, Feng Lei, Xiaojie Han, Linjie Chen, Shanshan Wu, Dapeng Ye

    UAV image acquisition and deep learning techniques have been widely used in field hydrological monitoring to meet the increasing data volume demand and refined quality. However, manual parameter training requires trial-and-error costs (T&E), and existing auto-trainings adapt to simple datasets and network structures, which is low practicality in unstructured environments, e.g., dry thermal valley environment (DTV). Therefore, this research combined a transfer learning (MTPI, maximum transfer potential index method) and an RL (the MTSA reinforcement learning, Multi-Thompson Sampling Algorithm) in dataset auto-augmentation and networks auto-training to reduce human experience and T&E. Firstly, to maximize the iteration speed and minimize the dataset consumption, the best iteration conditions (MTPI conditions) were derived with the improved MTPI method, which shows that subsequent iterations required only 2.30% dataset and 6.31% time cost. Then, the MTSA was improved under MTPI conditions (MTSA-MTPI) to auto-augmented datasets, and the results showed a 16.0% improvement in accuracy (human error) and a 20.9% reduction in standard error (T&E cost). Finally, the MTPI-MTSA was used for four networks auto-training (e.g., FCN, Seg-Net, U-Net, and Seg-Res-Net 50) and showed that the best Seg-Res-Net 50 gained 95.2% WPA (accuracy) and 90.9% WIoU. This study provided an effective auto-training method for complex vegetation information collection, which provides a reference for reducing the manual intervention of deep learning.


  • Open Access

    An Optimization Method of Deep Transfer Learning for Vegetation Segmentation under Rainy and Dry Season Differences in a Dry Thermal Valley

    by Yayong Chen, Beibei Zhou, Dapeng Ye, Lei Cui, Lei Feng, Xiaojie Han

    Deep learning networks might require re-training for different datasets, consuming significant manual labeling and training time. Transfer learning uses little new data and training time to enable pre-trained network segmentation in relevant scenarios (e.g., different vegetation images in rainy and dry seasons); however, existing transfer learning methods lack systematicity and controllability. So, an MTPI method (Maximum Transfer Potential Index method) was proposed to find the optimal conditions in data and feature quantity for transfer learning (MTPI conditions) in this study. The four pre-trained deep networks (Seg-Net (Semantic Segmentation Networks), FCN (Fully Convolutional Networks), Mobile net v2, and Res-Net 50 (Residual Network)) using the rainy season dataset showed that Res-Net 50 had the best accuracy with 93.58% and an WIoU (weight Intersection over Union) of 88.14%, most worthy to transfer training in vegetation segmentation. By obtaining each layer’s TPI performance (Transfer Potential Index) of the pre-trained Res-Net 50, the MTPI method results show that the 1000-TDS and 37-TP were estimated as the best training speed with the smallest dataset and a small error risk. The MTPI transfer learning results show 91.56% accuracy and 84.86% WIoU with 90% new dataset reduction and 90% iteration reduction, which is informative for deep networks in segmentation tasks between complex vegetation scenes.


  • Open Access

    Irrigation with Magnetized Water Alleviates the Harmful Effect of Saline–Alkaline Stress on Rice Seedlings

    by Changkun Ma, Qian Li, Zhaoxin Song, Lijun Su, Wanghai Tao, Beibei Zhou, Quanjiu Wang

    Saline–alkaline stress suppresses rice growth and threatens crop production. Despite substantial research on rice’s tolerance to saline–alkaline stress, fewer studies have examined the impact of magnetic water treatments on saline–alkaline-stressed rice plants. We explored the physiological and molecular mechanisms involved in saline–alkaline stress tolerance enhancement via irrigation with magnetized water using Nipponbare. The growth of Nipponbare plants was inhibited by saline–alkaline stress, but this inhibition was alleviated by irrigating the plants with magnetized water, as evidenced by greater plant height, biomass, chlorophyll content, photosynthetic rates, and root system in plants irrigated with magnetized water compared to those irrigated with non-magnetized water. Plants that were irrigated with magnetized water were able to acquire more total nitrogen. In addition, we proved that rice seedlings irrigated with magnetized water had a greater root NO3−-nitrogen concentration and root NH4+-nitrogen concentration than plants irrigated with non-magnetized water. These findings suggest that treatment with magnetized water could increase nitrogen uptake. To test this hypothesis, we analyzed the expression levels of genes involved in nitrogen acquisition. The expression levels of OsNRT1;1, OsNRT1;2, OsNRT2;1, OsAMT1;2, OsAMT2;1, OsAMT2;2, OsAMT2;3, OsAMT3;1, OsAMT3;2, and OsAMT3;3 were higher in plants exposed to magnetized water medium compared to those exposed to non-magnetized water media. We further demonstrated that treatment with magnetized water increases available nitrogen, NO3−-nitrogen content, and NH4+-nitrogen content in soil under saline–alkaline stress. Our results revealed that the increased resistance of rice seedlings to saline–alkaline stress may be attributable to a very effective nitrogen acquisition system enhanced by magnetized water.


  • Approximate analytical solution and parameter estimation for one‐dimensional horizontal absorption based on the van Genuchten–Mualem model

    by Lijun Su, Linfang Hou, Beibei Zhou, Yuyang Shan, Manli Duan, Yan Sun, Songrui Ning, Quanjiu Wang

    The van Genuchten–Mualem (VGM) model is used to describe the absorption of water by soil. The model is highly nonlinear, so developing a simple analytical or approximate method to predict the movement of soil water via the VGM model then estimating the model parameters is necessary. We present a approximate functional extremum solution (FES) for the problem of one‐dimensional horizontal absorption (1DHA) of water in soil under constant saturated boundary conditions and uniform initial conditions based on the least action and the variational principles. Simple linear relationships between the infiltration characteristics (distribution of soil water content, cumulative infiltration, infiltration rate, wetting front distance, and infiltration time) and the VGM model parameters were obtained. We developed a simple approximate analytical method for estimating the VGM model parameters by testing 1DHA. Numerical data simulated by HYDRUS‐1D were compared with the analytical results and used to verify our method's feasibility. The results indicated that the FES could accurately predict the infiltration characteristics, with relative errors between the approximate analytical and numerical solutions of 0.8 to 8.5% for the distribution of soil water content, 0.1 to 0.8% for cumulative infiltration, 1.1 to 6.4% for infiltration rate, and 0.1 to 6.9% for wetting front distance. Estimates of the parameters indicated that errors ranged from 3.2 to 28.3% for α (inverse of the mean pore diameter) and 0.4 to 8.0% for n (pore size distribution index). The proposed methods can be used to predict 1DHA and to estimate α and n in the VGM model.


  • Top Cited Papers

    Effects of Bacillus subtilis on carbon components and microbial functional metabolism during cow manure–straw composting

    by Manli Duan, Yuhua Zhang, Beibei Zhou, Zhenlun Qin, Junhu Wu, Quanjiu Wang, Yanan Yin


  • Open Access

    Comparison of the Effects of Polyacrylamide and Sodium Carboxymethylcellulose Application on Soil Water Infiltration in Sandy Loam Soils

    by Songrui Ning, Halidaimu Jumai, Quanjiu Wang, Beibei Zhou, Lijun Su, Yuyang Shan, Jihong Zhang

    Superabsorbent polymers have been used widely in agricultural production in arid and semi-arid regions to manage the soil water holding capacity. As the common water-retention polymers, the molecular weights, and structures of polyacrylamide (PAM) and sodium carboxymethylcellulose (CMC) are obviously different. Modified soil water management with polymers (i.e., PAM and CMC) has shown great promise for water conservation. Few researchers have reported the comparison of the effects of PAM and CMC on soil infiltration characteristics, especially in coarse-textured soils (i.e., sandy loam). In this research, two high-molecular polymers (PAM and CMC) were used to investigate the effects of polymers on soil water infiltration characteristics by laboratory experiment. The infiltration reduction effects of CMC treatments were more obvious than those of PAM treatments. With the applied rates of PAM (0.2–0.8 g/kg) and CMC (1–4 g/kg) increased, the processes of soil water infiltration were inhibited. The average infiltration time of CMC with different application rates is 1.85 times than that of PAM with different treatments. The mean wetting front distances of different application rates treatments of PAM and CMC were 22.20 and 19.23 cm. At the same application rate, applied CMC is more effective in reducing soil sorptivity than applied PAM in sandy loam soils. Moreover, the cost of application of CMC is lower than the cost of application of PAM. The mean economic inputs of PAM and CMC were 153.90 and 35.24 RMB/hm2. Therefore, CMC was selected and recommended as the suitable water retention agent in sandy loam soils.