|
| 基于MaxEnt模型提取撂荒耕地——以四川省武胜县为例 |
| Extraction of abandoned farmland based on MaxEnt model: A case study of Wusheng County, Sichuan Province |
| Received:July 31, 2021 |
| DOI：10.13254/j.jare.2021.0470 |
| 中文关键词: MaxEnt模型，撂荒耕地，遥感影像，时空分异，武胜县 |
| 英文关键词: MaxEnt model, abandoned farmland, remote sensing image, spatio-temporal differentiation, Wusheng County |
| 基金项目:国家自然科学基金项目（42071123）；四川省科技计划项目（2020YFG0033） |
| Author Name | Affiliation | E-mail | | LUO Yahong | School of Geography and Remote Sensing, Guangzhou University, Guangzhou 510006, China | | | GONG Jianzhou | School of Geography and Remote Sensing, Guangzhou University, Guangzhou 510006, China
South China Academy of Natural Resources Science and Technology, Guangzhou 510630, China | gongjzh66@126.com | | LI Tianxiang | Guangzhou Longterm Landscape Architecture Design Company Limited, Guangzhou 510520, China | | | HU Yueming | South China Academy of Natural Resources Science and Technology, Guangzhou 510630, China
College of Tropical Crops, Hainan University, Haikou 570228, China
College of Natural Resources and Environment, South China Agricultural University, Guangzhou 510642, China | |
|
| Hits: 1259 |
| Download times: 1326 |
| 中文摘要: |
| 在乡镇尺度厘清商品粮生产基地的撂荒耕地问题，对耕地保护和粮食安全具有重要意义。基于国产GF-1号遥感影像，耦合撂荒耕地的影响因子及影像波谱信息，以撂荒耕地问题较为突出的四川省武胜县为案例区域，探索应用MaxEnt模型提取常年性、季节性撂荒耕地信息的潜力，揭示撂荒耕地时空分异规律及其影响因素。结果表明，MaxEnt模型识别撂荒耕地的受试者工作特征曲线下面积（AUC）值均大于0.9，混淆矩阵总精度大于80%，季节性撂荒耕地面积与统计年鉴的相对误差不超过10%。受高程影响，常年性撂荒耕地主要集中分布于海拔超过300 m的丘陵山区，少数零星分散于嘉陵江两岸地势低缓的地区；季节性撂荒耕地各镇均普遍分布，局部呈片状分布特征。在2015—2018年研究时段内，常年性、季节性撂荒耕地面积和撂荒耕地总面积均保持平稳态势。研究认为，MaxEnt模型在提取撂荒耕地信息方面具有较大的应用潜力和优势；常年性与季节性撂荒耕地具有不同的空间分异特征，前者归因于海拔、交通及灌溉条件，后者归因于海拔、耕作半径和灌溉条件。研究丰富了基于遥感影像提取撂荒耕地信息的方法，增强了撂荒耕地时空分异特征与归因的认知，为乡村耕地合理利用与管理的实践提供理论支撑。 |
| 英文摘要: |
| Clarifying the problem of abandoned farmland at the township level is of great significance for the protection of farmland and food security. Based on domestic GF-1 remote sensing image, coupling the influence factors and image spectrum information of abandoned farmland and taking Wusheng County, where the problem of abandoned arable land is more prominent as the case area, the study is to explore the potential of using MaxEnt model to extract the information of perennial and seasonal abandoned farmland, reveal the spatial and temporal differentiation of abandoned farmland and its influencing factors. The results showed that the MaxEnt model had high accuracy and efficiency in the identification of abandoned farmland, which can be applied to extract the information of abandoned farmland. The relative error between seasonal abandoned farmland area and statistical yearbook was less than 10%. In 2018, the perennial abandoned farmland in Wusheng County was mainly distributed in the hilly and mountainous areas with an altitude of more than 300 m, and a few were scattered in the low-lying areas along the Jialing River. The seasonal abandoned farmland was generally distributed in each town, and the local distribution was patchy. During the 2015-2018, the area of perennial, seasonal and total abandoned farmland remained stable. This study suggested that MaxEnt model had great application potential and superiority in extracting abandoned farmland information. Perennials and seasonal abandoned farmland had different spatio-temporal differentiation patterns. The former was due to terrain, traffic and irrigation conditions, while the latter was due to farming radius and irrigation conditions.This study enriched the method of extracting abandoned farmland information based on remote sensing images, enhanced the cognition of the spatio-temporal differentiation patterns and attribution of abandoned farmland, and provided research support for the practice of rational use and management of rural farmland. |
| HTML
View Full Text
View/Add Comment Download reader |
| Close |
|
|
|
