College of Urban Economics and Public Administration, Capital University of Economics and Business, Beijing 100070, China.
College of Resources and Environmental Science, Wuhan University, Wuhan 430079, China.
Int J Environ Res Public Health. 2018 Jan 23;15(2):186. doi: 10.3390/ijerph15020186.
Thermal infrared remote sensing has become one of the main technology methods used for urban heat island research. When applying urban land surface temperature inversion of the thermal infrared band, problems with intensity level division arise because the method is subjective. However, this method is one of the few that performs heat island intensity level identification. This paper will build an intensity level identifier for an urban heat island, by using weak supervision and thought-based learning in an improved, restricted Boltzmann machine (RBM) model. The identifier automatically initializes the annotation and optimizes the model parameters sequentially until the target identifier is completed. The algorithm needs very little information about the weak labeling of the target training sample and generates an urban heat island intensity spatial distribution map. This study can provide reliable decision-making support for urban ecological planning and effective protection of urban ecological security. The experimental results showed the following: (1) The heat island effect in Wuhan is existent and intense. Heat island areas are widely distributed. The largest heat island area is in Wuhan, followed by the sub-green island. The total area encompassed by heat island and strong island levels accounts for 54.16% of the land in Wuhan. (2) Partially based on improved RBM identification, this method meets the research demands of determining the spatial distribution characteristics of the internal heat island effect; its identification accuracy is superior to that of comparable methods.
热红外遥感已成为城市热岛研究的主要技术方法之一。在应用热红外波段的城市地表温度反演时，由于方法具有主观性，会出现强度分级问题。但是，这种方法是为数不多的能够进行热岛强度分级识别的方法之一。本文将利用弱监督和基于思想的学习，在改进的受限玻尔兹曼机（RBM）模型中构建城市热岛的强度分级识别器。该识别器自动初始化标注，并依次优化模型参数，直到完成目标识别器。该算法仅需要少量有关目标训练样本的弱标注信息，并生成城市热岛强度的空间分布图。本研究可为城市生态规划和有效保护城市生态安全提供可靠的决策支持。实验结果表明：（1）武汉市存在明显的热岛效应，热岛分布广泛，其中最大的热岛区域在武汉市，其次是次绿岛。热岛和强岛级别的总覆盖面积占武汉市土地的 54.16%。（2）基于改进的 RBM 识别方法，该方法满足确定城市热岛内部效应空间分布特征的研究需求，其识别精度优于可比方法。
