Research On Estimation Method Of Building Volume Ratio In Zhongyuan District Of Zhengzhou City Based On GF-2 Remote Sensing Image

With the advancement of urbanization,the number of urban populations is increasing day by day,in stark contrast to limited land resources,and urban development and construction face enormous challenges.The plot ratio refers to the ratio of the total area of the building to the floor space.It is an important indicator to measure the intensity of urban land development and utilization,and has an important impact on urban housing prices and human settlements.The traditional method of calculating the floor area ratio is mainly based on manual field survey,which is time-consuming and labor-intensive,and has problems such as difficulty in updating data.With the development of remote sensing technology,the emergence of high-resolution remote sensing images provides a new data source for the extraction of urban building volume ratio.In this context,the use of remote sensing to extract the floor area ratio has the characteristics of real-time and wide coverage,which is convenient and fast,and greatly saves manpower and material resources.It has important practical significance for urban building information monitoring and urban management.This paper takes Zhongyuan District of Zhengzhou City as the research area,and uses the point of interest data and GF-2 remote sensing image as the main data source to divide the Zhongyuan District into multiple urban functional areas.The shadow length is established for each building area.The volumetric ratio inversion model such as method and shadow area method has completed the volume ratio extraction of buildings in Zhongyuan District,and analyzed the spatial distribution characteristics of volume ratio,which provides an important reference for urban building information monitoring and urban planning.The main findings are as follows:(1)This paper proposes a shadow extraction algorithm based on object-oriented multi-scale segmentation and random forest classification.According to the difference between shadow and other features,the algorithm constructs texture,shadow index,brightness and other features,effectively highlighting the shadow edge information,which is conducive to image segmentation.At the same time,accordingto the difference of different feature shapes,different segmentation scales are constructed to establish multiple image object layers.Using random forest classification algorithm,12 features such as shadow index,brightness,inverse moment,entropy and contrast of sample data are selected to construct a random forest classification model,and the model training is carried out to find the optimal number of generated trees(ntrees)and the maximum characteristics of random sampling.The number(mtry)finally realizes the classification of image features and extracts the precise shadow area,and the shadow extraction precision can reach more than 94%.(2)According to the complex and diverse characteristics of the buildings in the Central Plains,a variety of plot ratio extraction models are established.The Zhongyuan District has complex and diverse buildings.It only uses a plot ratio extraction model,and the error is large.Therefore,this paper statistically analyzes the density of interest points in each block,and divides the Central Plains into residential areas,industrial areas,science and education cultural areas,and public management.In the five districts of the district and the commercial district,different plot ratio estimation models are established according to the characteristics of the buildings in each functional zone.The commercial area and the public management area adopt the shadow length method,based on the geometric relationship between buildings,the sun and the shadow,establish a regression model between the shadow length and the height of the building;the urban residential area and the cultural and educational area are established by the shadow area method.The relationship model between the shaded area and the total area of the building;the suburban residential area and the industrial area adopt the building density method,thereby realizing the purpose of estimating the floor area ratio of the building by using remote sensing images.After testing,the accuracy of the plot ratio estimation is over 83%.(3)Based on the above research,this paper conducts an in-depth analysis of the volume ratio of the Central Plains.The overall distribution of the plot ratio in the Central Plains is uneven,showing the characteristics of high east and low west and high local center.The average plot ratios of the commercial area,science and education culture area,residential area,industrial area and green space square area are3.09,0.81,1.11,0.51,and 0.24,respectively.The analysis of the differentiation indexbetween different functional areas in the Central Plains area and the intra-regional differentiation index found that the difference between the regions accounted for70.25%,which is much larger than the intra-regional difference measure,which indicates the difference in volume ratio between the various functional areas of the city.Significantly,it is the main reason for the spatial difference in the volume ratio of the Central Plains.