|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/prediction-of-future-land-use-remote-sensing-and-gis-terrset/
课程评论：没有评论
**课程名称：** 利用GIS预测未来土地利用（TerrSet - CA Markov - ArcGIS） **课程概述：** 本课程将通过TerrSet（原IDRISI）软件，运用CA Markov模型，结合ArcGIS的数据准备和Erdas的部分功能，展示机器学习在土地利用变化预测中的实际应用。本课程**不涉及任何编程**，所有使用的软件**非开源**，学员需自行解决软件获取问题。 课程要求学员**具备制作土地利用图的基础技能**。课程将从零开始，深入讲解未来的土地利用预测过程，重点关注影响土地利用变化的各种驱动因素（Parameters），例如： * **城市边界：** 新建居民区通常倾向于靠近现有城市边界。 * **道路和地形：** 早期城市扩张倾向于沿河流和公路。 * **可转化区域：** 分析农田、森林等区域转化为城市用地的可能性。 * **限制性区域：** 模型需要设定规则，避免将水体（河流、湖泊、水库）等区域纳入城市用地预测。 此外，课程还将涵盖模型**学习精度和输出精度的调整与评估**。例如，课程中实现了42%和67%的学习精度，以及89%的预测精度。学员在实际操作中，精度可能因计算机和数据而异，但重要的是**理解和关注最终的预测结果**。课程强调，100%的学习精度反而可能意味着模型设置不当。因此，学员需认真观看并理解每一段视频，然后再运行模型。 **课程级别：** 专家级。建议在报名**前观看免费预览视频**。IDRISI Taiga 用户也可遵循相同步骤。 **课程特点：** 90%的实践操作，10%的理论讲解。 鼓励学员在问答环节积极提问。
In this course you will see Machine learning in Action using readymade land Change model Terrset (formerly IDRISI ). This course used Terrset Software with CA Markov method to predict future landuse ArcGIS is used to prepare data. Erdas also used for some task. No coding is used.All software used in this course are NOT Open Source. You need to manage software. You must know to prepare landuse maps rest of things covered in this course from scratch. Future prediction of landuse depends on number of drivers/Parameters. Drives means forces which decide how the future urban area will look. It includes many drives like, old city boundary because new settlement will be constructed near to old city boundary. Roads and relief are also one of factors, because first roads near city covered by settlement. On another side how, much possibility at different location on agriculture site that can be convert to urban. Similarly, forest cover also. We also need to avoid some landuse classed like water, river, lake or reservoir never convert to urban. So, we need to setup our model in such a way so that it avoids water. After setting accuracy of learning and output accuracy also matters. We also need to modify it. In this course we have achieved learning accuracy of 42%, and 67% in two different runs. But 89% accuracy we have achieved in predicted landuse. Learning and prediction accuracy is different on computer to computer and data to data. While running you will receive more or less accuracy then this course. But focus on your output results. If Learning accuracy was 100% then it also wrong. So, see and understand each video carefully. Then run you model. You must see free preview video before enrolling this course. Because this is Expert level course. Note: Who having IDRISI Taiga They can also follow same steps. This course covers 90% Practical and 10% Theory. Don't hesitate to ask me Questions in QA Session.
