Research on the Development of “Ghost City” Based on Night Light Data: Taking Sichuan Province as an Example

Due to the improvement of urbanization level and unreasonable development in China, many cities have begun to appear as “Ghost City”; that is, the phenomenon of high vacancy rate and low occupancy rate in urban areas. The emergence of this phenomenon will seriously affect the healthy development of cities. Therefore, the monitoring and analysis of the urban “Ghost City” index is of great significance to urban population and urban construction. This paper uses Landsat8 remote sensing image data, night light remote sensing image data, and resident population data of various cities and counties in Sichuan Province to calculate the “Ghost City” index, and obtains the “Ghost City” Index of prefecture-level cities and county-level cities in Sichuan Province. Based on the judgment criteria of the Ghost City, the calculation results show that three cities have become “Ghost City” within the research scope; eight cities have a tendency to develop into “Ghost City”; the rest of the cities have developed well and there has not been a “Ghost City” phenomenon. According to this conclusion, this paper studies the similarities and differences between the cities that are the ghost towns or the cities with the development trend of ghost towns, analyzes the reasons for the formation of ghost towns and makes suggestions for the urbanization of Sichuan Province, and provides reference for the direction of urban development and rational planning.

urbanization process has gradually accelerated, and the demand for housing has also risen sharply. The National Urban Land Use Data Summary Results Analysis Report [1] mentioned that the national urban land area increased by 1.65 million hectares in 2009-2014, with an average annual growth rate of 4.2%, indicating that urban construction land is increasing in a large amount. The development and unreasonable development of the city led to the emergence of empty cities or ghost cities, that is, the phenomenon of high housing vacancy rate and low occupancy rate [2].
Due to the insufficient monitoring of the dynamic changes of urban population information, the phenomenon of "Ghost City" has appeared one after another. The emergence of this phenomenon has seriously affected the development of the city and restricted the advancement of urbanization, resulting in the waste of land resources and weakening urban operational efficiency. The criterion for "Ghost City" is that the ratio of urban population to built-up area is less than 0.5 or slightly higher than 0.5 (the standard is based on the land occupation standard from the Ministry of Housing and Urban-Rural Development, the built-up area per square kilometer holds 10,000 people), this criterion known as the "Ghost City" Index. This study takes Sichuan Province as an example, combined with the 2013 Landsat8 OLI_TIRS and DMSP/OLS night-time remote sensing image data, through the fitting of the urban population and the extraction of the built-up area, to achieve the city-level and county-level cities in Sichuan Province. The analysis and monitoring of the "Ghost City" index, and through the analysis of the reasons for the formation of the ghost city phenomenon, propose corresponding countermeasures, and provide the basis for urban development and planning.

Research Scope
Sichuan Province has complex landforms and has four types of landforms: mountains, hills, plains and plateaus. Therefore, this paper selects Sichuan Province as the research scope and provides reference for the study of ghost cities in other cities in China. Sichuan governs 18 prefecture-level cities, 3 autonomous prefectures, 17 county-level cities, 108 counties, and 4 autonomous counties. In this study, only 18 prefecture-level cities and 17 county-level cities were selected as research areas analysis ( Figure 1) and monitoring of the "Ghost City" index in Sichuan Province.

DMSP/OLS Data Processing
In order to obtain the remote sensing image data of the study area, the obtained DMSP/OLS luminous remote sensing image data is tailored according to the administrative boundary vector data of Sichuan Province, and the DMSP data of various cities and county-level cities in Sichuan Province are obtained ( Figure 2).

Extraction of Urban Built-up Areas of Landsat8 OLI_TIRS Data
The Landsat data of the studied cities were analyzed, and the land use data including the use of construction land, forest land and water bodies were obtained.  The important basis for assessing the level of urban development is urban construction land. Researchers at home and abroad have made a lot of research on the extraction of urban construction land. Among them, Xu Hanqiu [3] proposed to use the IBI building land index to extract urban built-up areas from remote sensing images. Accuracy can reach more than 96%. The IBI building land index can be constructed by the three indexes of normalized building index NDBI, vegetation cover index NDVI and improved normalized water body index MNDWI. Calculation formula as follows: Among them, B3, B4, B5 and B6 are the green band, the red band, the nearinfrared band, and the SWIR1 band of the Landsat8 OLI_TIRS image.
The IBI building land index formula is as follows: Substituting the Equation (1) Due to space limitations, the processing results of all prefecture-level cities and county-level cities cannot be displayed. Only the construction land and built-up areas extracted by the IBI index of Mianyang City are displayed here, as shown in Figure 3 and

Research Ideas and Methods
The research process of this paper for the ghost city index ( Figure 5) is as follows:    the study of the ghost city index method, and the data will be continuously improved in future research.

Population Spatialization
The administrative boundary vector data of Sichuan Province is overlapped with the night light remote sensing image, and the DMSP/OLS data of each city and county level city is obtained through the cropping in the GIS data management tool, and the data of night light intensity greater than 6 is obtained through attribute extraction, and after extraction, the data is extracted. The light intensity distribution is shown in Figure 6.    According to Figure 8, the matching of the city's light intensity value with the resident population has a good effect. Therefore, the population quantity information extracted by real-time remote sensing data is scientific and accurate. The Note: DN is the sum of the light gray values of the city and county level cities.

"Ghost City" Index
Ghost City is an urban disease in the process of urban development due to unrestricted expansion, lack of rational planning and unbalanced development, and the urban housing vacancy rate is too high and the occupancy rate is too low. In  attention to the development of "healthy cities". The existence of the "Ghost City" phenomenon is inevitably inconsistent with the trend of urban healthy development. The "Ghost City" index is an important indicator to measure whether the city has become a "Ghost City" [4]. Therefore, the monitoring of the "Ghost City Index" is particularly important.
According to the built-up area of prefecture-level cities and county-level cities in Sichuan Province extracted from the previous article and the urban population obtained according to the fitted regression equation, the "Ghost City" index of prefecture-level cities and county-level cities in Sichuan Province is calculated ( Table 2). The calculation formula is as follows: Note: Index is the "Ghost City" index of the Prefecture-level city and County-level cities; Num is the urban population; S is the built-up area.

Analysis of Calculation Results
According to the Ghost Town Index obtained above, 35 cities in Sichuan Province (18 prefecture-level cities and 17 county-level cities) can be divided into three categories: There are 3 cities with an index less than 0.5, including Zigong, Luzhou and Ya'an. Such cities have become ghost towns, and their urban development has serious problems, and there has been an imbalance between urban population and urbanization. Therefore, it is necessary to implement new urbanization and realize ghosting urbanization through rational use of land resources and planning.
There are 8 cities with indices ranging from 0.5 to 0.8, including Dazhou, Nanchong, Yibin, Suining, Guanghan, Wanyuan, Xichang and Jiangyou. Although this type of city has not become a ghost town, its urbanization still has problems. It also needs to provide a more rational urbanization development strategy for the city in terms of policies and planning.
There are 24 cities with indices greater than 0.

The Reason for the Formation of Ghost Towns
According to the previous analysis of the ghost city index, the ghost towns in Sichuan Province are mainly distributed in the southeast, and the cities with ghost town development trends are mainly distributed in the northeast. There is basically no ghost town phenomenon in central Sichuan. The central part of Sichuan Province is dominated by cities such as Chengdu and Mianyang, and Chengdu, as the capital of Sichuan Province, is therefore more attractive to people, so there is no ghost city phenomenon. According to a large number of literature research, many scholars also pay more attention to the ghost city phenomenon. Among them, Nie Xiangyu [5] and Duan Lufeng [6] have elaborated on the types of ghost towns. The types are mainly disaster-type ghost cities, fading ghost cities, and planning advanced type ghost cities, planning lag ghost cities and real estate ghost cities. Among them, the disaster-type ghost town is mainly caused by the city's lack of popularity, which leads to becoming an empty city. The decline-type ghost town is caused by the depletion of resources in the process of urban development, resulting in the loss of population, thus forming a ghost city. Planning advanced ghost towns is Due to unreasonable planning, far exceeding the demand of urban planning population, many construction land is idle. Planning a lag-type ghost town is because the planning cannot meet the needs of the urban population, the planning function is too single or the public service facilities and infrastructure are not perfect, resulting in urban population outflow. Real estate-type ghost town is due to the emergence of real estate speculation, resulting in the expansion of the real estate market, making the urban housing occupancy rate too low, vacancy rate is too high, thus forming a ghost town.
According to the study of the similarities and differences between cities (Table   3), the cities that formed the ghost city phenomenon and the development  trend of ghost towns in this study are not in line with the disaster-type and the decline-type ghost city. Most of the cities (such as Luzhou, Ya'an, etc.) are mainly tourism development. Due to the excessive development of tourism, most of the buildings or service facilities in urban construction are aimed at tourism. In order to pursue the interests of the tourism industry, the large-scale construction of the tourism industry has led to a reduction in the utilization rate of urban land, thus forming a ghost town. In addition to this, there are also ghost towns created by cities in order to promote development, large-scale construction of new districts, and new districts that are not attractive to the population and enterprises.

Conclusions and Suggestions
In this study, the 2013 DMSP/OLS night light remote sensing image and Land- sources, the lower the ghost city index; the higher the utilization rate of urban land resources, the higher the ghost city index. "Ghost Town" is always closely linked with the urbanization of the Great Leap Forward and over-exploitation of real estate. It is a subjective "making city" movement that lacks rational cognition and scientific considerations [7]. Therefore, the study of the Ghost Town Index is conducive to the monitoring of the urbanization of the city, which is conducive to the healthy and rapid development of each city.
The existence of ghost towns will result in low land resource utilization and waste of land, and will hinder the healthy and rapid development of cities. Therefore, how to rationally urbanize is a key issue for urban development. Combined with the previous analysis of the Sichuan Ghost Town Index and the reasons for the formation of ghost towns, the paper puts forward suggestions on the future of ghosting and rational urbanization in Sichuan. 1) Government level: In the process of urban development, the stage of growth of urban development is overemphasized, while the ups and downs of urban development may be neglected [8]. Therefore, in the process of development, scientific decision-making should be made to rationally plan the situation of the city. The distribution law of urban population density is declining from the traditional central area to the suburbs, so the intensity of human activities in the newly developed areas of the city is significantly lower than that of the urban old city [9]. In the process of urbanization, the construction of the new district must be gradual and gradual, and it is necessary to strengthen the supervision and management of urban planning to prevent the emergence of unorderly spreading [10] and lead to the formation of "empty cities" and "ghost cities".
2) Planner level: Each city has its own unique culture and resources. In the planning, it should avoid the "Same Imagines of the City" and all construction methods cannot be copied. Planning and design should focus on the development of traditional culture with urban characteristics, strengthen residents' sense of identity with the city, and enhance the vitality of the city.
In summary, ghost towns will bring various environmental problems such as land waste and waste of resources. Cities such as Ya'an and Zhangzhou should be committed to revitalizing "ghost towns". Therefore, each city should implement scientific and rational urban planning. When planning and constructing, it should be realistic and conform to local conditions and integrate local culture to enhance the city's attractiveness and vitality.