KJKai Jian

Kai Jian

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Kai Jian has a total of 15 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Forest carbon stock estimation in Wuyishan National Park using multisource data and Bayesian-optimized residual dropout network

    by Kai Jian, Guiying Li, Liyin Wang, Dengsheng Lu

    1. Accurate assessment of forest carbon stock (FCS) and its spatial distribution is fundamental for evaluating the ecological capacity of national parks, formulating conservation strategies for pro...

  • Open Access

    Modeling Forest Carbon Stock Based on Sample Plots and UAV Lidar Data from Multiple Sites and Examining Its Vertical Characteristics in Wuyishan National Park

    by Kai Jian, Dengsheng Lu, Guiying Li

    The accurate estimation of forest carbon stocks with remote sensing technologies helps reveal the spatial patterns of forest carbon stocks within national parks, but the limited number of sample plots in one site often results in difficulty in developing robust estimation models. This study employed a Bayesian hierarchical model to estimate forest carbon stock based on data from 193 sample plots collected across 37 UAV (unmanned aerial vehicle) Lidar sites. The developed model was employed to predict the carbon stock distribution in 17 Lidar sites within Wuyishan National Park (WNP). Then, the carbon stock characteristics along vertical zones of vegetation distribution (VZsVD) were examined. The results showed an overall coefficient of determination (R2) of 0.84 for forest carbon stock estimation across four regions, with a root mean square error (RMSE) of 12.09 t/ha. Within WNP, the overall R2 was 0.73, with specific values of 0.83 for broadleaf forests, 0.61 for mixed forests, 0.53 for Masson pine forests, and 0.46 for Chinese fir forests. Despite variations in R2, the relative RMSE (rRMSE) averaged 20.15%, ranging from 10.83% to 23.57%. The average carbon stock was 52.15 t/ha. Forest diversity and structural complexity emerged as key factors influencing the vertical distribution of carbon stocks. Regions with complex and diverse forest types exhibited higher and more evenly distributed carbon stocks. Chinese fir and Masson pine showed higher carbon stocks in low-altitude regions (350–850 m) than other vegetation types. In medium- to high-elevation regions (1350–1600 m), the carbon stocks of mixed forest and broadleaf forests remained relatively stable. Conversely, coniferous forests at high altitudes (above 1600 m) had lower carbon stocks due to extreme climatic and terrain conditions. This study provided a comprehensive analysis of carbon stock distribution across different VZsVD in WNP, offering valuable insights for enhancing the management of national parks.


  • Open Access

    Improving Forest Canopy Height Mapping in Wuyishan National Park Through Calibration of ZiYuan-3 Stereo Imagery Using Limited Unmanned Aerial Vehicle LiDAR Data

    by Kai Jian, Dengsheng Lu, Yagang Lu, Guiying Li

    Forest canopy height (FCH) is a critical parameter for forest management and ecosystem modeling, but there is a lack of accurate FCH distribution in large areas. To address this issue, this study selected Wuyishan National Park in China as a case study to explore the calibration method for mapping FCH in a complex subtropical mountainous region based on ZiYuan-3 (ZY3) stereo imagery and limited Unmanned Aerial Vehicle (UAV) LiDAR data. Pearson’s correlation analysis, Categorical Boosting (CatBoost) feature importance analysis, and causal effect analysis were used to examine major factors causing extraction errors of digital surface model (DSM) data from ZY3 stereo imagery. Different machine learning algorithms were compared and used to calibrate the DSM and FCH results. The results indicate that the DSM extraction accuracy based on ZY3 stereo imagery is primarily influenced by slope aspect, elevation, and vegetation characteristics. These influences were particularly notable in areas with a complex topography and dense vegetation coverage. A Bayesian-optimized CatBoost model with directly calibrating the original FCH (the difference between the DSM from ZY3 and high-precision digital elevation model (DEM) data) demonstrated the best prediction performance. This model produced the FCH map at a 4 m spatial resolution, the root mean square error (RMSE) was reduced from 6.47 m based on initial stereo imagery to 3.99 m after calibration, and the relative RMSE (rRMSE) was reduced from 36.52% to 22.53%. The study demonstrates the feasibility of using ZY3 imagery for regional forest canopy height mapping and confirms the superior performance of using the CatBoost algorithm in enhancing FCH calibration accuracy. These findings provide valuable insights into the multidimensional impacts of key environmental factors on FCH extraction, supporting precise forest monitoring and carbon stock assessment in complex terrains in subtropical regions.


  • Open Access

    Vertical Characteristics of Vegetation Distribution in Wuyishan National Park Based on Multi-Source High-Resolution Remotely Sensed Data

    by Yongpeng Ye, Dengsheng Lu, Zuohang Wu, Kuo Liao, Mingxing Zhou, Kai Jian, Dengqiu Li

    Identifying vertical characteristics of mountainous vegetation distribution is necessary for studying the ecological environment quality and biodiversity and for evaluating its responses to climate change. However, producing fine vegetation distribution in a complex mountainous area remains a huge challenge. This study developed a framework based on multi-source high-resolution satellite images to strengthen the understanding of vertical features of vegetation distribution. We fused GaoFen-6 and Sentinel-2 data to produce 2 m multispectral data, combined with ALOS PALSAR digital elevation model (DEM) data, and used an object-based method to extract variables for establishing a classification model. The spatial distribution of vegetation types in Wuyishan National Park (WNP) was then obtained using a hierarchical random forest classifier. The characteristics of different vegetation types along the elevation gradient and their distribution patterns under different human protection levels were finally examined. The results show that (1) An overall accuracy of 87.11% and a Kappa coefficient of 0.85 for vegetation classification was achieved. (2) WNP exhibits obviously vertical differentiation of vegetation types, showing four compound dominant zone groups and five dominant belts. (3) The composition of vegetation types in the scenic area differs significantly from other regions. The proportions of Masson pine and Chinese fir exhibit a noticeably decreasing trend as the distance increases away from roads, while the changes in broadleaf forest and bamboo forest are less pronounced.