可占用准租金视角下农村集体建设用地流转租赁制改进研究——基于江苏省宜兴市的调查

探索如何高效率地解决集体建设用地流转租赁制中可能存在的"敲竹杠"问题，有助于优化我国集体建设用地租赁流转方式。企业生产设备的资产专用性高，产生的可占用准租金高，使企业面临被村集体"敲竹杠"的风险。江苏宜兴市通过流转双方互惠协议和声誉机制等自动实施契约联结，且租赁交易受到第三方干预机制——市政府文件中明确履约保证的约束，共同避免了村集体的"敲竹杠"行为。为促进集体建设用地租赁制健康发展，应借鉴宜兴"长租短约制"，延长租赁期限，稳定租赁关系，避免土地所有者通过抽回土地的威胁来"敲竹杠"。同时，政府应发挥市场规范者的角色，建立清晰的违约责任惩处机制。     

Global assessment of land vulnerability to water erosion on a ½° by ½° grid

A simple methodology is presented for assessing the risk of water erosion at the global level. It uses a ½° latitude by ½° longitude soil database, developed at ISRIC, and auxiliary databases on climate and land cover with a similar spatial resolution. Area estimates are presented for (1) susceptible areas, as determined by rainfall erosivity, topography and soil erodibility, and (2) vulnerable areas as further determined by the pressure of current land use. Model output for vulnerability is evaluated against observed data on severity of soil degradation by water as presented on the map of human-induced soil degradation (GLASOD). Cross-tabulation of the vulnerability and GLASOD subsets gave a significant Cramer's correlation coefficient of 0·72 (P ≪ 0·005). Thus a fair geographic agreement was observed between the grid cells considered vulnerable to water erosion, under current conditions of land cover, and regions in which water erosion occurs currently. The qualitative model can serve to raise awareness on issues of soil degradation by water at the global level by identifying regions at risk, where more detailed studies are needed. However, it does not provide any information on the actual rate of erosion at the field scale, nor on the associated decrease in crop productivity and biodiversity. The study of productivity changes associated with water erosion, at different scales, is currently an important topic on ISRIC's research agenda.     

基于激光雷达与深度相机融合的SLAM算法

针对单一传感器地图构建时存在环境表征不充分，无法为移动机器人自主导航提供完整环境地图等问题，本文通过将激光雷达与深度相机获取的环境信息进行互补融合，构建出更完整精确的栅格地图。首先，对传统ORB-SLAM2算法进行增强，使其具备稠密点云地图构建、八叉树地图构建以及栅格地图构建等功能。其次，为验证增强后ORB-SLAM2算法的性能，在fr1＿desk1数据集和真实场景下进行测试，数据显示增强后ORB-SLAM2算法绝对位姿误差降低52.2%,相机跟踪轨迹增长14.7%,定位更加精准。然后，D435i型深度相机采用增强型ORB-SLAM2算法，激光雷达采用的Gmapping-Slam算法，按照贝叶斯估计的规则进行互补融合构建全局栅格地图。最后，搭建实验平台进行验证，并分别与深度相机和激光雷达2个传感器建图效果进行对比。实验结果表明，本文融合算法对周围障碍物的识别能力更强，可获取更完整的环境信息，地图构建更加清晰精确，满足移动机器人导航与路径规划的需要。