摘要
水深或水深点是海图的重要内容，水深综合是海图制图综合的重要方面，也是海图制图综合的难题之一。本文在分析海图水深综合要求的基础上，研究了水深综合的神经元网络的网络结构与作业策略、网络参数、网络学习、网络实现，并给出了实验结果。
As a main part of nautical chart cartographic generalization, the generalization of soundings is also one of the bottlenecks in the way of automatic chart generalization. In the article, the design of sounding generalization neural network was discussed, along with the analysis on the network's overall structure, the working methods, network factors, learning. To solve the problem of dealing with both the spatial and attributive factors at the same time to select soundings, a new set of operations is designed based a practical problem solving method called “Hierarchical Information Structure”. The experimental result is shown upon a protocol system.
出处
《测绘学报》
EI
CSCD
北大核心
1999年第4期335-339,共5页
Acta Geodaetica et Cartographica Sinica
关键词
海图
人工神经元网络
网络结构
水深综合
chart generalization
artificial neural network
network structure
network factors
network learning
