Introduction

Urban heritage promotes the transformation of resource-based cities

Heritage research shows a trend of regionalization and integration from a global perspective. It shifts from emphasizing individual heritage to group heritage and a holistic environment. At the same time, heritage's connotation and protection scope have expanded [1], such as Cultural Landscapes, Heritage Corridor, Heritage Trails, etc. [2]. It is necessary to comprehensively study the widely existing heritage objects at the macro scale. The World Heritage Center (WHC) selection criteria have evolved, now prioritizing the integrity of Outstanding Universal Value (OUV) at a broader regional scale [3]. Additionally, the research advocates for a shift from a singular building focus to a comprehensive, regional-wide perspective in managing urban cultural heritage. It promotes using the Historic Urban Landscapes (HUL) approach to navigate change and address pressing challenges in urban heritage conservation, such as balancing preservation with development, acknowledging cultural diversity, and managing the impacts of tourism [4, 5].

Recently, the development of resource-based cities has shifted from incremental expansion to high-quality stock upgrading in the new Chinese urbanisation era. Resource-based cities face problems such as resources excessive dependence, unbalanced industrial structure, and stalling urban development [6]. Urban Heritage is crucial to developing resource-based cities and achieving high-quality urbanisation for sustainable development [7,8,9]. Also, heritage is vital in stimulating spatial vitality in the urban environment and enhancing cultural tourism in urban areas [10].

Urban vitality can well reflect the urban environmental quality [11, 12]. As a fundamental urban management tool, heritage management effectively activates the urban environment by increasing its spatial vitality [13]. In 2003, ICCROM (International Centre for the Study of the Preservation and Restoration of Cultural Property) launched a "living heritage sites" plan, which defined the living heritage approach (LHA) as a management and protection method based on "community continuity and variability" [14], and transferred the protection practice from material authenticity to heritage attached to living places [15]. The ultimate goal of heritage management is to creatively integrate heritage into civil life and establish a sustainable urban development model [16]. Cultural heritage is essential for revitalizing historical cities from the social sustainability perspective [17, 18].

However, the protection and activation of cultural heritage are usually focused on architectural heritage rather than the historic cultural landscape [19, 20]. Few successful cases of activating urban development through architectural heritage protection led to a "heritage fever" [21]. Excessive development results in resource mismatch due to the pursuit of economic growth. In addition, decision-makers and designers do not comprehensively understand the heritage's value [22]. The consideration of heritage conservation and redevelopment is still in isolation from the urban context and rarely put into the urban network for overall cognitive analysis and management.

Heritage corridor is a concept and an approach developed in the United States to protect and manage heritage regionally and globally [23]. Heritage corridors are now defined as linear or spatially connected areas that preserve and integrate historical, cultural, and natural heritage assets while promoting sustainable urban development [24]. This approach emphasizes the significance of cultural and natural resources value of heritage [25], and constructing a heritage corridor usually requires the coordination of culture and environment [26]. Heritage can be better recognized through urban morphology analysis as part of the urban components [27, 28]. By establishing a heritage corridor, the holistic consideration of heritage and the surrounding urban built environment elements could be objectively evaluated and analyzed in the context of an urban spatial network [29]. In heritage corridors, environmental spatial evaluation examines the interaction between physical space and heritage preservation. Urban vitality—including economic activity, social interaction, and accessibility—is influenced by the presence and quality of heritage corridors [30]. Effective evaluation can enhance urban vitality by creating cultural spaces, strengthening the integration between heritage corridors and urban environments [31]. Conducting environmental spatial assessments within heritage corridors is essential for balancing cultural preservation with urban development. This approach uniquely integrates quantitative and qualitative heritage aspects with urban dynamics, ensuring that development does not compromise heritage value while promoting sustainable urban growth.

Quantitative analysis of spatial quality based on multi-source data: ESM

Recent research highlights a critical gap in integrating cultural heritage protection with urban development. While heritage conservation often focuses on architectural elements, there is a notable lack of coordinated strategies that address both cultural and environmental factors comprehensively [19,20,21]. Studies reveal that existing approaches frequently treat heritage conservation in isolation, failing to consider how heritage interacts with broader urban contexts and spatial networks. This disconnection results in missed opportunities for optimizing heritage value within urban environments.

This investigation into heritage corridors has shown promise in bridging this gap, yet they still face challenges. Although heritage corridors are intended to integrate historical, cultural, and natural assets while promoting urban vitality, there is insufficient understanding of how these elements function together within urban morphology [32, 33]. Current studies used the following elements as the index to measure urban vitality: street accessibility, construction intensity, architectural form, and function mixing degree [34]. Most principles of urban design can be summarised into these elements. For example, as the elements for promoting urban vitality mentioned by Jacobs [35], "adequate traffic density" and "short streets" support and promote accessibility, respectively, and "adequate traffic" will be related to the construction intensity and architectural form. In addition, "the functional mix" and "multiple architectural ages" indicate the mixing degree of functions. In terms of urban vitality, Jan Gehl proposes that "integration rather than separation" and "convergence rather than dispersion" both require appropriate street accessibility, and "invitation rather than exclusion" and "openness rather than closure" contribute to forming a vibrant street life [36]. Multiple classical theories have been verified through quantitative environmental measurement [37,38,39]. The measurement method based on multi-source data provides a new paradigm for current urban scientific research.

New technology emerged, and appropriate quantitative urban network analysis methods expand heritage conservation notions and arouse multiple impact spatial factors that need to be considered [40]. For example, as scholars mentioned, urban facilities [41] and the accessibility of public transport [42] have a significant influence on urban vitality and residents’ satisfaction. Therefore, it is necessary to include these factors in a holistic "cultural heritage construction and protection" decision-making process. At the same time, the open data of map providers [43], as well as the support of ArcGIS spatial network analysis [44], make it possible to implement this idea [45, 46].

In addition, various digital heritage technologies for information collecting (Photogrammetry, 3D scanning) [47,48,49] and analyzing (GIS, BIM, and deep learning) [50,51,52,53] also provide tools for presenting heritage values and contribute to the activation of urban spatial environment. With the emergence of street spatial identification technologies through AI-based image semantic segmentation, the Pyramid Scene Parsing Network (PSPNet) is one of these technologies that use the pyramid pooling module to build a scene analysis network [54,55,56]. It uses the pyramid pooling module to construct a scene analysis network, which can integrate the context of different regions, thus improving the ability to obtain global information. Many scholars have realized the integrated analysis of greening and other elements from the perspective of human eyes through the analysis of PSPNet and used it to measure space quality [54, 55, 57]. It generally believes that green vision has the greatest impact on the quality of space, the richness of facilities, and sidewalks are also major concerns of pedestrians [58]. Also, streets with solid wall interfaces are less attractive to people, but streets with grey spaces and interfaces are attractive [59].

This research proposes extending heritage assessment frameworks, such as the HUL and the LHA, which focus on managing and protecting heritage through "continuity and variability." By integrating these approaches with advanced spatial analysis and multi-source data techniques, we introduce a comprehensive Environment Spatial Model (ESM). This model leverages emerging heritage spatial typologies, landscape measurement methods, and diverse data acquisition approaches to provide a holistic assessment of heritage corridors. Additionally, verifying the ESM through PSPNet analysis could help explore proper and efficient heritage environment functions, layout, and management strategies. The study can also cultivate a research methodological reference for the heritage corridor environment in domestic cities along linear river space, offering robust support for revitalizing resource-based cities for high-quality urbanization. It aims to address the following research questions:

  1. (1)

    How to establish a framework based on ESM to evaluate heritage corridors systematically?

  2. (2)

    How are the spatial environments of heritage corridors classified, and what are their distribution characteristics?

  3. (3)

    How can heritage corridor upgrading strategies based on ESM be applied to promote resource-based urban revitalization?

The research purposes are as follows: Establish an ESM method to evaluate heritage environment characteristics, take the FMD as an example, collect urban environmental data around the local heritage site, and guide the development of heritage areas. Based on the results of qualitative cognition with PSPNet analysis and quantitative judgment of the quality of the local heritage corridor environment [60], strategies for optimizing and improving heritage corridors are put forward by portraying and classifying heritage with different characteristics.

Data and methods

Research site

This research selected Fengfeng Mining District (FMD) in Handan City, Hebei Province, as the research area, which is a resource-based city for its coal and steel industries (Fig. 1). According to administrative division data (Table 4), FMD covers an area of 353 square kilometers and comprises 9 towns, 1 township, and 1 subdistrict area. With a population of 530,000 and a 2023 GDP of 22.13 billion RMB, it represents a resource-dependent, mixed industrial and agricultural zone where urban and rural landscapes intersect. Due to historical changes and war damage, the spectacular scenery of the ancient capital of Handan has gone. However, Handan's limited remaining cultural relics seriously lack conservation in the historical environment due to urban sprawl, and the contradiction between fragmentation and atomization of the relics is prominent. The main heritage of such cities consists of small and medium-sized cultural relics scattered throughout the city. Thus, little data is available for reference, and the data sample is limited. Therefore, quantitative analysis based on multi-source data is selected. It aims to provide methods and suggestions for activating heritage at the urban scale by collecting and analyzing relevant indicators of heritage corridors with the ESM.

Fig. 1
figure 1

Study area

Full size image

The study offers preliminary research results for constructing the "Fuyang River Cultural Belt" in Handan. The theoretical basis for choosing FMD stems from the Resource-Based City Transformation Theory [61]. Sustainable development in resource-dependent cities requires industrial diversification through structural adjustment and technological innovation, reducing reliance on a single resource [62]. Heritage plays an active role in urban regeneration during this transition. For instance, Japan’s policy-driven reemployment of miners in coal-depleted areas serves as a practical example of heritage-led regeneration [63]. FMD exemplifies the challenges and opportunities in regenerating cities with significant industrial heritage. The area faces unique difficulties due to its reliance on declining industrial activities and weak economic and innovation bases. This case study offers insights into using heritage resources to drive urban regeneration and high-quality development in similar resource-dependent cities. The potential findings can guide similar regions facing industrial heritage challenges. The methods and findings aim to be applicable across different contexts through comparative analysis. We discuss how these results can be applied to other resource-based cities and regions in China and globally, particularly by drawing analogies with other industrial and heritage-rich areas. Although the study focuses on FMD, the findings hold significant relevance for other resource-based cities facing similar challenges, emphasizing the broader applicability of the research.

Along with the Fuyang River, with a variety of cultural heritage scattered on both sides, FMD, as a resource-based industrial and mining city is gradually changing its local industry from mining to cultural tourism. Considering the heritage resource base or industrial transformation background, this paper selects a series of heritage spots and their sounding environment along the Fuyang River (Fig. 1) in FMD as the research object is more appropriate. It uses the " heritage corridor " as a comprehensive synthesis to restructure these loosely distributed heritages better for further evaluating the urban vitality. With the Fuyang River as the clue, the study integrates the 22 major heritage sites in FMD (Table 1). It explores practical protection and utilization strategies based on the quality of heritage space.

Table 1 List of 22 main heritage information in FMD
Full size table

When investigating and classifying the heritage of the Fuyang River in FMD, it was found that there were few grotto temples and stone carvings, and no mural heritage. Therefore, referring to the classification of immovable cultural relics in the Implementation Plan of the Third National Cultural Relics Census and Relevant Standards and Specifications (hereinafter referred to as the "Standards and Specifications of the Third National Cultural Relics Census") in 2007, the above-defined heritage classification has been refined again. Therefore, all heritages along the Fuyang River in FMD are archived in six categories: ancient sites, ancient tombs, ancient buildings, grotto temples and stone carvings, important modern historical sites, and representative buildings, and others (Table 1).

The 22 heritage sites collected and screened in the study are distributed on both sides of the Fuyang River, mainly ancient buildings and sites of the Ming (1368–1644) and Qing (1636–1912) Dynasties. The Yangjiaopu site, with abundant historical and cultural heritage, dates back to the Han Dynasty (202 B.C.–220 A.D.). In addition, in the investigation, we found that local heritage was authorized as provincial or national cultural relics protection units. Still, the conservation practice by the official department was unsatisfactory, and uncontrolled tourists damaged the local heritage sites, so the improvement of heritage protection was urgent. In such areas, the research value of heritage is high, but the protection and development system does not match and lacks pertinence. Our research means great significance and value for developing and activating heritage.

Conceptual framework

Environmental spatial evaluation is crucial to the transformation and development of resource-based cities for its assistance in activating heritage corridors [4, 13, 24]. The research framework is mainly based on quantitative analysis with multivariate data. Although many studies involve urban big data and network research [10, 54, 64, 65], few articles have been involved in the environmental spatial evaluation under the background of resource-based cities. Based on the above ideas, this study proposes a heritage ESM based on Traffic, Facility, and Ecology and conducts an assessment.

The process begins by obtaining the geographic coordinates of heritage sites along the heritage corridor, alongside multi-source spatial data from the city, as shown in Fig. 2, including urban roads, POI (Point of Interest), AOI (Area of Interest), remote sensing satellite data and other geospatial big data such as rivers, administrative boundaries, buffer zones, etc. Moreover, the multi-source spatial data and satellite imagery covering the heritage corridor have been cross-verified to maintain consistency in the spatial elements. Then we apply these data to the analysis of eight dimensions: Road Accessibility (RA), Public Station (PS), Public Facility (PF), Facility Accessibility (FA), Facility Diversity (FD), Distance from the Factory (DF), Green Land Contact (GL), and Green Coverage (GC). Next, heritage images are collected to capture a comprehensive visual representation of these sites within the urban context. A wearable camera mitigates discrepancies caused by varying photo angles, ensuring a fixed perspective and consistent view. The ESM analysis results are superimposed and then divided into four main categories of heritage through cluster analysis and verified via PSPNet-based image semantic segmentation. Finally, the heritage activation strategy is explored based on data analysis results and site clustering.

Fig. 2
figure 2

The conceptual framework of Heritage Corridor research

Full size image

The research framework (Fig. 2) of this paper is attached as follows:

The first part shows the data collection and research methods. Currently, scholars have more quantitative research on evaluating urban environment conditions. Most scholars' research on environment space is based on social network data, POI, and other urban primary and remote sensing data [66,67,68]. Remote sensing data (Landsat 8, using 30 m resolution true-color images synthesized from the 432 bands) were primarily used to capture various spatial elements of the site, such as the road network, water bodies, and green coverage. These data formed the foundation for the spatial analysis map. However, for less developed areas, social media and communication information data are rare, so these data are used as auxiliary data. It uses road data, POI data, remote sensing data with high popularity as the primary analysis data, and the rest of the river, AOI, and administrative data as auxiliary and expression. These data are mainly obtained from digital map suppliers and field surveys.

The second part evaluates the urban vitality based on the heritage ESM framework we established, including three factors: traffic, facilities, and ecology. At the "Traffic" level, scholars believe that road accessibility significantly impacts the environment [69], and public transport is one of the main factors affecting the environment [66]. At the "facility" level, POI data correspond to functions or facilities in cities or villages. Some scholars believe that POI density substantially affects the comprehensive environment [65], and mixed land use directly drives comprehensive urban dynamism [70,71,72,73]. Therefore, we can illustrate the environment's performance at the functional space level from PF, FA, and FD. In terms of "ecology", the ecological environment is a significant factor that significantly affects the environment of the city [74]. Sound ecology helps improve urban space's livability and is necessary to stimulate heritage space's effectiveness.

In the third part, PSPNet was been used to analyze the spatial semantics of the environmental image of heritage. Heritage images are captured to obtain a comprehensive visual representation within the urban context. To prevent variations caused by differences in camera angles, wearable cameras were employed, securing a fixed shooting position to maintain consistent perspectives. We verified it by identifying the spatial scene, including the building, tree, grass, chair, lamp, board, lane, sky, bridge, river, and other elements in the image, and generating their percentage [75]. Then prioritize greening, facilities, sidewalks, and others in research results [76]. The study is based on the proportion of 5, 4, 3, 2, and 1 as the space score. In this way, the ESM assessment results could be verified by the PSPNet image analysis.

The fourth part provides strategies with several categories based on the second part of the site space measurement results. The second part divides the quantitative measurement results and their superposition results into several main categories through K-means clustering, for its capacity to generate distinct clusters [77]. As there are several main heritage classes in the heritage corridor, the quantitative measurement provides a more objective perspective on the heritage classes. The relevant heritage activation strategies are generated based on heritage classification and evaluation results.

Research methodology

geographical analysis via GIS

This paper built a heritage ESM from three aspects: Traffic, Facility, and Ecology, and expanded it through eight elements: RA, PS, PF, FA, FD, DF, GL, and GC. See Table 2 for the indicators and their meanings.

Table 2 Quantification calculation method for heritage ESM impact factors designed in our framework
Full size table

Traffic is evaluated with two indicators, RA and PS. sDNA(Spatial Design Network Analysis) is the world-leading 2D and 3D spatial network analysis software for GIS (Geographic information system), CAD, Command Line & Python using industry-standard network representation [78, 79]. According to indicators of Network Quantity Penalized for Distance Analysis (NQPDAn), sDNA analysis results, the accessibility of roads around heritage can be reflected. \({N}_{PS}\) index refers to the number of bus stops within 1 km around heritage, which reflects the improvement of public transport services at the site.

PF, FA, and FD are three indicators used to reflect the facility environment. \({K}_{poi}\) and \({N}_{PF}\) is the statistics of POI points 1 km around heritage, in contrast, \({K}_{poi}\) emphasises the concept of "quantity", which is the POI quantity, \({N}_{PF}\) is the quantity calculated according to the relevant standard [80] and a specific weight (the weight of catering, shopping, medical care, education and transportation is 0.5, the weight of life services, sports and leisure, science and education and culture is 0.3, and the weight of others is 0.2), with more emphasis on the "quality" of facilities; By contrast \(-{\Sigma }_{i=1}^{k}\frac{{n}_{i}}{N}\mathit{ln}\frac{{n}_{i}}{N}i\) is the embodiment of facility diversity, which is the expression of Shannon Wiener index, where k is the number of functional categories of POI points; \({n}_{i}\) is the number of type \(i\) functions; \(N\) is the total number of all POI facilities in the selected area. The larger the FD index, the better the diversity.

Environmental quality and green space status are important manifestations of ecology. D is Distance, Distance from the factory (\({D}_{Factory}\)) is the distance from the nearest factory. GL takes heritage center as the benchmark, R = 1000 m, R = 2000 m, R = 12500 m. It counts the area of green space accessible to heritage pedestrians, cyclists, and cyclists in three dimensions. It calculates by superposition with a weight of 0.5:0.3:0.2 to obtain each heritage's green space contact value. GC represents the greening degree around heritage, \({S}_{green}\) is the area of green patches counted on remote sensing satellite images within 1 km around heritage, \({S}_{land}\) is the area of the whole region, and the ratio of the two is the value of GC.

For the results of heritage ESM evaluation, the normalized processing conducted and the relevant factors obtained, assign values 1, 2, 3, 4, and 5 to the results respectively, and then conduct multi-factor superposition. The values obtained will be used as the comprehensive heritage ESM evaluation results. In this study, PSPNet was used to perform semantic analysis on images collected at a fixed interval of 30 m from the space of 10 randomly selected sites, and to score them according to the same criterion, rank them, and compare and verify the analysis on the ESM results.

Although the ESM framework has been relatively systematic and complete, it is also necessary to use qualitative means as verification [81,82,83]. The pre-trained model (ResNet101) and dilated/atrous convolution of the site image are used to extract the feature map. The extracted feature map is 1/8 of the input size. The feature map gets the fused feature with overall information through the Pyramid Pooling Module. The up-sampling is concatenated with the feature map before pooling. Finally, the final output is obtained through a convolution layer. Then, images were scored according to the area occupied by different recognition elements. A K-means cluster analysis through SPSS then taken according to the evaluation results and various indicators and divided heritages into four classes, A, B, C, and D, which summarises each characteristic and then puts forward adaptive design strategies and policies for activation of various heritages [84].

Image semantic segmentation via PSPNet

PSPNet, a powerful machine learning tool for scene parsing and semantic segmentation, precisely identifies various objects and regions within images, which this study uses to segment selected heritage images and compare the results with ESM analysis [54,55,56]. The segmentation of elements such as the sky, buildings, and roads in the sample images helps quantify the current state of the heritage site. By distinguishing these components, we can assess the spatial quality and environmental context of the heritage, providing a clearer understanding of how these elements interact within the urban landscape. The specific principle is shown in Fig. 3.

Fig. 3
figure 3

Machine learning principle diagram

Full size image

Data source

According to the framework, geospatial big data is the foundation for the ESM evaluation of heritage corridors. The heritage image data and basic information data are used to revise and improve the collected data through fieldwork surveys and interviews. The purpose, content, and source of these data are shown in Table 3.

Table 3 The input datasets used in the research framework
Full size table

According to the evaluation framework shown in Fig. 2, the geospatial big data and primary information data described above are obtained (Table 3). The road network data is derived from the OpenStreetMap (OSM) road network data in 2020, which is the main data used for road accessibility analysis, and is obtained through satellite map images and field surveys; POI and AOI data are obtained through the Gaode (高德) Map Open Platform. POI includes large, medium, and small categories. Public transport station data and factory location data are obtained by classification screening. AOI data mainly includes spatial distribution data of urban and rural parks, forest farms, etc. Remote sensing satellite data is mainly obtained through Google Earth software, which is mainly used for additional correction of other data and analysis and extraction of green coverage data. Through field investigation, we selected 22 typical heritage sites from the local area, created a 1 km buffer zone through ArcGIS for research, and obtained administrative divisions and other data mainly through the official website of the Ministry of Civil Affairs. In addition, in the survey, we took a large number of heritage images, each around the 30 m*30 m scale. This study used 75 images of 10 selected heritage (see Table 4).

Table 4 Heritage data acquisition lists and sources
Full size table

Result

Quantitative evaluation results of heritage through ESM

The quantitative results of the eight indicators of heritage in the Fengfeng Mining Area were generated through the ESM assessment (Table 5 and Fig. 4). The evaluation results are divided into eight levels by the natural segment classification method. In the traffic evaluation, the road axis network is analyzed using space syntax tools, and it shows that the west side of the town and the Gushan area have the strongest accessibility. The more westward the direction, the lower the accessibility. From the perspective of the accessibility distribution of bus stops around the local heritage site, the new urban area (Linshui Town) of FMD and the old site of Fengfeng Power Plant has the highest accessibility. The heritage sites distribution map decreases from southeast to northwest, and is bounded by the Gushan mountain range composed of Gushan Mountain and Shenzhuan Mountain. The east side of the mountain range is better than the west side; The bus accessibility of Pengcheng Town near the west of the mountain range is general, while that of Yijing Town and Hecun Town in the northeast is poor, which is in line with the actual situation that the bus operation line extends from the southeast to the northwest of the city. The fieldwork research shows that there are few bus stops and lines near the heritage sites in the upper reaches of the Fuyang River, villages and towns, and Yijing Town. The blue area in Fig. 4 shows that the bus accessibility is relatively poor, which proves the qualitative judgment on the bus accessibility in this area.

Table 5 Heritage Evaluation Results based on ESM
Full size table
Fig. 4
figure 4

Around the heritage site a Accessibility of road b Distribution number of bus stops c Distribution number of facilities d Distribution of facility accessibility e Diversity distribution of facilities f Distance from factories g Contact distribution of green space h Green space coverage

Full size image

In the facility evaluation, the number of facilities around heritage is still the highest near the old site of Fengfeng Power Plant and the ancient tunnel. The Linshui Town area (Fig. 1), where the old site of Fengfeng Power Plant and the ancient tunnel is located, is generally higher. At the same time, the west side of Gushan Mountain is significantly lower. Midwest Pengcheng Town and Yijing Town are upstream until fewer facilities are distributed among the villages and towns, and there is no significant difference. The accessibility distribution of facilities around heritage is roughly consistent with the distribution of the number of facilities. The former site of Fengfeng Power Plant is the highest near the site. The east side of the power plant is higher, the west side is significantly lower, and the facilities in other areas are less distributed. The diversity distribution of facilities around the heritage site is more homogeneous than the above accumulation analysis results. As a whole, the diversity of facilities around the heritage site in Linshui Town (the new urban area) of FMD is the strongest, followed by Pengcheng Town (the old urban area), Yijing Town, and Hecun Town have a medium diversity of facilities, east of Yijing Town has weak diversity of facilities, and south of Yijing Town have the lowest diversity of facilities.

For the ecology evaluation, the adverse environmental exposure in the middle of the river is low. The community gathering place has a good facility environment, and many green spaces can be accessed easily. Such areas are mainly located in cities and towns with convenient transportation and the upstream regions of rivers. At the same time, the former site of Tonger Mine, situated in the east of the upstream river and the Longwangmiao in Yijing, located in the east of the middle river, is also close to the factory, so the adverse environmental exposure is high. It is highest near Fengyue Pass and National Wetland Park, lower in the middle reaches, and the gap between upstream and downstream is not apparent.

The indicators obtained from each heritage are normalized according to Sect. 3.2.1 (Table 5) to obtain its overall ESM. Through K-means cluster analysis, 22 heritage sites are categorized into Class A, B, C and D. In Table 5, Sum is the result of the superposition of RA, PS, PF, FA, FD, DF, GL and GC, and Rank is the classification of heritage according to the superposition result.

The PSPNet measurement

The PSPNet's measurement results of 10 sampled (relatively representative) heritage images are shown in Table 6. The samples were selected according to the identified category based on ESM analysis results, aiming to cover a diverse range of heritage spaces for PSPNet analysis. The ESM offers a comprehensive measurement of tangible and intangible elements, while PSPNet primarily reflects the quantifiable aspects rooted in the HUL approach. Thus, PSPNet evaluates the current state of individual heritage sites, whereas ESM provides a more balanced approach, integrating preservation with revitalization, and offering adaptive strategies tailored to heritage conservation.

Table 6 PSPNet verification evaluation results
Full size table

The results indicate that Cross Street Coupon Pavilion has the lowest score (0.05), and Yangjiaopu Site has the highest score (1.18). The ESM assessment and PSPNet results are assigned in high and low levels using the color red to blue respectively. The results show that the ESM assessment results are basically consistent with the PSPNet assessment results, that is, the ESM assessment results could be verified their reliability through the PSPNet assessment results.

The assessment results of the three sites do not conform to this rule. First, Fengyue Pass, whose PSPNet assessment results are in the middle, has almost the highest ESM results. On the contrary, Jinci Temple Site and Cizhou Kiln Temple have high PSPNet results with low ESM results. Through the judgment of the fieldwork investigation to these three sites, we found that Fengyue Pass is located in Fengfeng Town, close to the Fuyang River. Still, its surroundings and interior are old and monotonous. In contrast, the site of Jinci Temple is located at the foot of the mountain in the middle of the town, with a good ecological environment and a wide view. The deviation of PSPNet assessment results is that its measurement range is limited, ignoring the influences of surrounding environments, while the ESM made a comprehensive judgment from the reverse side of facilities, transportation, environment, etc., leading to some differences in results. This ESM approach illustrates potential strategies for heritage preservation and revitalization. For example, the differences observed in places like Jade Emperor Pavilion, Jinci Temple Site, and Cizhou Kiln Temple stem from variations in the intrinsic quality of the heritage and its surrounding environment. These discrepancies highlight how ESM's integration with urban space offers a more nuanced evaluation. At the scale of heritage corridors, this method can significantly contribute to the high-quality transformation of resource-based cities.

In Fig. 5, for the overall heritage ESM distribution of the heritage corridor, the heritages with a high score are mainly concentrated near the town, the heritages in the north of the middle section have the lowest score, and the heritages in the northwest end has a general rating. Therefore, the environment's spatial evaluation level of heritage is related to the heritage status quo and location. The heritages in the town can fully use some facilities brought by the town to urban residents. In contrast, the area in the north of the middle section is in rural areas, where the supporting facilities are relatively weak, affecting the heritage rating. The heritage in this area is D-Class.

Fig. 5
figure 5

Evaluation Results of Heritage Environment Spatial Evaluation (The number in the figure is the heritage number of Table 1, which is placed above the legend)

Full size image

Discussion

Establishment of heritage ESM and verified with PSPNet

Through the heritage ESM described above, the local heritage can be assessed, and the heritage environment data can be presented more objectively, making up for the subjective experience-oriented human judgment. And with the analysis of the PSPNet assistant, the ESM evaluation results could be verified and justified coordinately. This assessment-verification system could eliminate the model deviation and provide an integrated guiding basement. Then, according to the different levels of evaluation results, more correlated interventions can be paid to the conservation and development of the heritage corridor. Improve the opportunities for entertainment and economic development, and eliminate the dispersion, confusion, and fragmentation of the existing space. Using this methodology to perform and analyze an environmental space could enhance the possibility of studying heritage corridors and proposing tailored-made strategies for heritage conservation.

In the analysis, it was found that heritage had a great relationship with geographical spatial distribution. That is, heritage evaluation results were higher when it was located in the belt of the city and the northwest residential gathering areas, and heritage with the worst score was distributed in rural areas. Based on this issue, we argue that activating heritages should be consistent with its physical and geographical context. The heritage located in developed urban areas should be given priority to development, and its heritage revitalization strategy should be integrated with the surrounding environment to improve its holistic regional quality. The reasons are, First, heritage around the town has better environmental conditions, which could save less time and costs in the activation process and serve more residents after its upgrading. Secondly, heritage in rural areas has always been regarded as an obstacle to urban development because the infrastructure construction is backward, and it is difficult to have adequate accessibility, and its overdevelopment could damage the local landscape.

Clustering of heritage and the improvement application

The four types of heritages clustered according to K-means have the following characteristics:

According to the environment spatial evaluation, A-Class heritage represents the highest potential for revitalizing its surrounding region through heritage management. The commonalities in A-Class heritage lie in improved accessibility and environmental quality, which facilitate the integration of these sites into diverse urban contexts. Consequently, revitalization strategies for A-Class heritage will focus on openness, scenographic appeal, and experiential qualities. Uniquely, sites such as Fengyue Pass, Xiangtang Mountain Grottoes, and Zhifang Yuhuang Pavilion in the urban core can extend heritage revitalization to surrounding areas, creating interconnected corridors and enhancing accessibility and commercial experience. In contrast, dispersed sites on the city’s periphery, like Yangjiaopu Site and Yijing Dragon King Temple, could be integrated into a comprehensive countryside heritage park, combining with rural cultural facilities to create public cultural spaces.

B-Class heritage also has a high potential for revitalizing its surrounding region but with limited defects at different levels. The distribution of B-Class heritage is scattered and has the same common features. For B-Class heritage, common traits include well-preserved heritage and robust surrounding infrastructure, offering the potential to enhance the cultural and social fabric of the city. Uniquely, the Ancient Tunnel in the city center could leverage digital experiences, using its unique spatial characteristics for immersive lighting and digital scenarios, potentially becoming a popular attraction. In suburban village clusters, such as Cross Street Coupon Pavilion and Li Yi's Tomb of Tang Dynasty, integrating educational and recreational facilities could establish cultural and educational hubs, promoting local culture and improving connectivity with Category A sites, fostering regional synergy.

C-Class heritage generally has good traffic conditions but with obvious defects in its facilities and ecological environment quality, and the overall value of its environment spatial evaluation is lower than that of A-Class and B-Class. For C-Class heritage, common traits are mainly concentrated in the villages on the north side, where the heritage is well preserved and mostly of the temple type. Uniquely, the Former Site of Fengfeng Power Plant in the city center offers expansive space with potential for industrial heritage tourism. Meanwhile, heritage sites like Cui Fujun Temple, Martyrs Cemetery, and Jiaying Linghui Temple in He Village, currently open for visits, could leverage their geographical advantage to become a cultural heritage destination. Therefore, for C-Class heritage, priority should be given to improving the quality of the ecology and facilities in its region.

D-Class heritage has no significant advantages, and the overall environment spatial evaluation is low. Located in remote urban–rural areas, safeguarding D-Class heritage is necessary, and the heritage administrative department should consider the protection of the heritage with regular inspection. D-Class heritage consists mainly of infrastructure sites, mostly scattered in rural areas and currently neglected, requiring protection and selective revitalization. Uniquely, the Mountain Anti-Japanese Tunnel Site in the countryside could be enhanced with historical and educational functions for commemorative events. In the city, the Cizhou Kiln (Yandian) Site could strengthen connections with other heritage areas, creating a cohesive experience group, enhancing city walks and storytelling, and exploring further revitalization opportunities. In addition, the narrative and spiritual value of D-Class heritage should be considered in combination with its historic cultural landscape, which is more relevant to the research trend of environmental justice [85].

Conclusions

To protect and develop the heritage corridor in a resource-based city, this research integrates the complex historical, cultural and ecological environmental issues on a regional scale. More diversified and comprehensive research methods are proposed in this research, promoting more precise protection and development strategies. In this way, it solves the problems of fragmentation of natural and cultural landscapes and fragmentation of heritage protection and makes heritage revitalization more integrated. This paper analyses the heritage corridor protection and development research in the Fengfeng Mining District, a resource-based city transitioning to a cultural revitalization economy. It presents a study that aims to evaluate the heritage environment by analyzing its surrounding spatial elements through the Environmental Spatial Model (ESM) with GIS-based data and verification through AI-based image semantic segmentation analysis. Additionally, this study incorporates property indicators, such as PF, FA, and other factors, into the GIS analysis, aligning with the financial toolset recommended by the HUL approach. These integrated methodologies offer development strategies to safeguard heritage, enhance its potential, and contribute to the revitalization of holistic urban–rural areas.

Quantitative urban environment research has changed the current study and cognitive way of people in cities. At the same time, ArcGIS-based spatial network analysis tools have matured, the threshold of urban research has been lowered, and urban environment spatial evaluation with multi-sourced data has been developed. This study effectively combines urban multivariate big data and spatial network measurement methods to propose a holistic framework for measuring the environment of heritage corridors. By analyzing 22 heritage sites in the FMD, we found that heritage site assessment results in variations and heterogeneity with the geographical distribution. The heritage characteristics and the holistic spatial quality define the heritage categories. The results show that the ESM assessment results are basically consistent with the AI-based image semantic segmentation results; that is, the results of ESM assessment could be verified. Then, the existing heritage is clustered through the K-means clustering method, and four types of heritage with typical characteristics are obtained. Last but not least, the four types of heritage and targeted restoration and optimization strategies are proposed based on ESM analysis. This study offers strategies for the future heritage spatial development and evolution of resource-based cities.

This paper also has some limitations. The analysis framework of this paper is place-based, which gives more consideration to the urban environment. Due to the limited access to heritage financial data and social network data, people-based consideration is relatively lacking, and the ESM does not cover holistic financial and geospatial data. Additionally, using images is only one way to serve as a verification tool to integrate further the spatial relationships between heritage corridors and the factors affecting corridor attributes to ensure a more comprehensive assessment; the original regional living state impact needs further study after activating some heritage through the above heritage activation strategy. Future research could incorporate subjective evaluations from residents and tourists, providing a deeper understanding of heritage impacts. Integrating quantitative tools and humanistic and financial data would lead to a more comprehensive, people-centered assessment of heritage corridors’ conservation and development, refining strategies for effective management and urban revitalization.