人工神经网络在城市湿地生态环境质量评价中的应用

Application of artificial neural network in assessment of urban wetland eco-environmental quality.

松北区是哈尔滨城市空间"跨江"发展战略的重点开发区域。该文以哈尔滨松北区城市湿地为研究对象,应用人工神经网络建立生态环境质量评价的BP网络模型,选择6个因子作为松北湿地生态环境质量的评价指标,用训练好的BP网络进行评价。结果表明,松北湿地整体生态环境质量的综合评价结果为勉强合格(0.611 6);人工神经网络用于生态环境质量的评价结果与环境质量实况相符,为松北城市湿地的健康发展提供了理论依据;同时在应用人工神经网络模型对湿地生态环境质量评价中,针对不同地区的生态系统可适当地增加网络的隐节点或引层数,以提高神经网络的学习能力及训练效果。

The north shore of Harbin City is a major development zone in carrying out the programs of enlarging urban areas along river regions in Harbin.In this study,urban wetland of the north shore of Harbin was the study object and the assessment model of eco-environmental quality by BP artificial neural network was established.Six factors were selected as evaluation indices and well-trained networks used to assess eco- environmental quality.The results indicate that overall evaluation of wetland eco-environmental quality in the north shore of Harbin was barely enough qualified (0.611 6),and evaluation result by artificial neural network accorded with the fact.These data provided the theorectical basis for healthy development of urban wetland.At the same time,according to ecosystem from different regions,evaluation of wetland eco-environmental quality by artificial neural network model should increase network secret nodes or layer numbers to improve learning ability and training effect.