*Article* **The Land Use Mapping Techniques (Including the Areas Used by Pedestrians) Based on Low-Level Aerial Imagery**

## **Maciej Smaczy ´nski \*, Beata Medy ´nska-Gulij and Łukasz Halik**

Department of Cartography and Geomatics, Faculty of Geographical and Geological Sciences, Adam Mickiewicz University in Pozna ´n, 61-712 Pozna ´n, Poland; bmg@amu.edu.pl (B.M.-G.); lhalik@amu.edu.pl (Ł.H.)

**\*** Correspondence: maciej.smaczynski@amu.edu.pl

Received: 9 November 2020; Accepted: 12 December 2020; Published: 16 December 2020

**Abstract:** Traditionally, chorochromatic maps with a qualitative measurement level are used for land use presentations. Along with the use of UAV (Unmanned Aerial Vehicles), it became possible to register dynamic phenomena in a small space. We analyze the application of qualitative and quantitative mapping methods to visualize land use in a dynamic context thanks to cyclically obtained UAV imaging. The aim of the research is to produce thematic maps showing the actual land use of the small area urbanized by pedestrians. The research was based on low-level aerial imagery that recorded the movement of pedestrians in the research area. Additionally, based on the observation of pedestrian movement, researchers pointed out the areas of land that pedestrians used incorrectly. For this purpose, the author will present his own concept of the point-to-polygon transformation of pedestrians' representation. The research was an opportunity to demonstrate suitable mapping techniques to effectively convey the information on land use by pedestrians. The results allowed the authors of this article to draw conclusions on the choice of suitable mapping techniques during the process of thematic land use map design and to specify further areas for research.

**Keywords:** thematic map; mapping techniques transformation; land use; UAV; geometric representation; pedestrian; chorochromatic maps; dot map; heat map

#### **1. Introduction**

The employment of Unmanned Aerial Vehicles (UAV) for obtaining data about changes in land use in the form of images occurs in small areas in which various changes can be observed. Thanks to their capability to position themselves at a constant height and remain stationary for a specific period, multirotor UAVs can provide opportunities to observe the phenomenon that occurs in each area. The ability to start air raids every couple of minutes over the area of several hectares allows one to record people moving around. In our research, we have touched upon the problem of producing thematic maps of the land cover, illustrating official land use, including places whereby the use of which has been changed by pedestrians.

Methods of obtaining UAV data require creating a digital elevation model (DEM) and a point cloud based on the images obtained [1–3]. The actual reflection of geometry of the recorded area or object, including defining its actual location, is particularly interesting. To do so, researchers focus on GCPs that serve as reference points used in the aerial triangulation of imagery [4–6]. Using geodetic techniques of measurement allows one to have accurate coordinate values and accurate heights of the photogrammetric models [7–11].

It becomes crucial to properly adjust mapping methods of presenting raster data transformations from images into qualitative and quantitative vector data. We can see the necessity to develop well-known methods of cartographic presentation for visualizing data obtained from UAVs. Obtaining static data based on images is relatively simple, but the observation of and recording changes in dynamic phenomena in the form of images requires a particular approach [12]. Recording space used by pedestrians outside designated sidewalks is one of such phenomena. We assume that actual land cover also includes also the areas occupied by pedestrians outside designated sidewalks, e.g. worn shortcut paths or areas along sidewalks used when the sidewalk is too narrow for all pedestrians.

The aim of the research on cartographic visualization is to obtain highly effective spatial information. The effectiveness of cartographic visualisations connect to the creation of maps that enable the simple and unequivocal reading of features of geographical phenomena [13]. The effectiveness of visualisation is defined as both the efficiency and effectiveness of the communication of information [14]. It is directly proportional to the quantity of information correctly received and is inversely proportional to the time used to obtain it [15].

The increase in cartographic effectiveness significantly improves the usefulness of maps [16] by suitable mapping methods and graphic variables, with which we can capture and demonstrate changes occurring in the area [17–19]. The choice of the mapping technique is key to thematic map design. Mapping techniques are specific methods applied to the cartographic presentation of spatial phenomena and the relations existing there between. The following factors are decisive for the effectiveness of the obtained map: measurement level, geometric form of objects, and graphic variables. The classification of attributes is effected on specific measurement levels by the assignment of numbers or categories to objects, which is tantamount to the presentation of the relations existing between these objects by suitably selected measurement scales. Each map comprises a set of graphical elements that present spatial data [20]. First and foremost, general geographical and topographical maps provide nominal information, and to a lesser extent, quantitative data. Several mapping methods may be used in any one map. Quantitative data are presented applying a statistical approach, and the end-product of the utilization of this specific mapping method are statistical maps. The geometrical form of objects requires the adoption of an appropriate mapping method. Point phenomena may be represented to emphasise qualitative or quantitative attributes.

For the representation of areal phenomena, use is made of the following: chorochromatic maps, choropleth map, dasymetric maps, areal cartodiagrams, scope maps, isoline maps and statistical surfaces. Apart from the division into qualitative and quantitative data, it is worth mentioning the regular or irregular nature of the distribution of areal phenomena, which frequently entails the selection of basic fields for data [21]. Satellite or UAV images, which show types of land cover for individual grids, are called chorochromatic graticule maps, where each grid cell has a different colour.

The dot method is considered as a variant of the point symbol method, but on the dot map, a small circle becomes the sign, the size of which enables the more precise location of the phenomenon. The most logical method of location entails introducing the assumption that one dot presents one object, but in practice a specific value is given to a single dot; this is known as the dot ratio [22].

Heat maps (specific variant of the choropleth map), with various color scales used for areas with different intensity of points that represent magnitude of a given phenomenon, are particularly beneficial for visualizations and the exploration of large quantitative data sets [23–26].

The combination and transformations of mapping methods of point and areal phenomena require appropriate graphical procedures that make it possible to maintain the legibility of the cartographic visualisation effect, i.e., enabling the unequivocal interpretation of spatial relations.

The main objective of our work was to work out thematic maps based on UAV imagery that would present the actual use of the small, urbanized area, including places for which pedestrians changed the form of use. In this research, we pinpointed a few intermediary goals:


land use by pedestrians.

the residents of the nearby housing estates.

• To use the dot map, buffer map, chorochromatic map and the so-called *heat map* to present quantitative data by means of the area method, and for cartographic presentation of the actual land use by pedestrians. • To adjust mapping methods to the point-to-polygon transformation of pedestrians' representation. • To use the dot map, buffer map, chorochromatic map and the so-called *heat map* to

*ISPRS Int. J. Geo-Inf.* **2020**, *9*, x FOR PEER REVIEW 3 of 17

• To demonstrate the effectiveness of the thematic maps created, presenting the actual land use by pedestrians. present quantitative data by means of the area method, and for cartographic presentation of the actual land use by pedestrians. • To demonstrate the effectiveness of the thematic maps created, presenting the actual

## **2. Study Area**

The inclusion of pedestrian movement in the identification of the actual land use has become the main determinant in the process of pinpointing conditions the research area should be characterized by. We have assumed that the use of the research area by pedestrians should be as intense as possible. For the purposes of the research, we concluded that urbanized areas would be most intensely used by pedestrians. A relatively dense and diverse form of urbanized areas, as well as the nature of the phenomenon analyzed, and technology employed for observation, should help one reduce the size of the research area. In this study, we assumed that the maximum research area, depending on the specificity of construction of selected areas, would be no larger than 10 ha. The construction of the research area should be balanced and include both buildings and structures, such as corridors (roads, sidewalks, parking lots), objects of the so-called small architecture and green spaces. The study was conducted in the western part of Poland, the Greater Poland province, in the city of Pozna´n. In accordance with the standards adopted in terms of the research area characteristics, we selected the area that was a part of the campus of Adam Mickiewicz University in Pozna´n. The area is 9 ha and met all the selected criteria (Figure 1). A chosen fragment is one of the parts most intensely used by pedestrians, as it is located between the main part of the campus and the railway station. The area selected for the research is used daily by thousands of students and university employees, as well as the residents of the nearby housing estates. **2. Study Area**  The inclusion of pedestrian movement in the identification of the actual land use has become the main determinant in the process of pinpointing conditions the research area should be characterized by. We have assumed that the use of the research area by pedestrians should be as intense as possible. For the purposes of the research, we concluded that urbanized areas would be most intensely used by pedestrians. A relatively dense and diverse form of urbanized areas, as well as the nature of the phenomenon analyzed, and technology employed for observation, should help one reduce the size of the research area. In this study, we assumed that the maximum research area, depending on the specificity of construction of selected areas, would be no larger than 10 ha. The construction of the research area should be balanced and include both buildings and structures, such as corridors (roads, sidewalks, parking lots), objects of the so-called small architecture and green spaces. The study was conducted in the western part of Poland, the Greater Poland province, in the city of Poznań. In accordance with the standards adopted in terms of the research area characteristics, we selected the area that was a part of the campus of Adam Mickiewicz University in Poznań. The area is 9 ha and met all the selected criteria (Figure 1). A chosen fragment is one of the parts most intensely used by pedestrians, as it is located between the main part of the campus and the railway station. The area selected for the research is used daily by thousands of students and university employees, as well as

**Figure 1.** Location of the studied area (background layer orthophotomap from national geoportal www.geoportal.gov.pl). **Figure 1.** Location of the studied area (background layer orthophotomap from national geoportal www.geoportal.gov.pl).

Then, we analyzed the research area in terms of transport infrastructure (Figure 2). Our priority was to mark out areas for pedestrians, i.e., sidewalks, routes for pedestrians and cyclists, parking lots and access roads. These elements were included in the research as areas in which pedestrian movement was allowed (Figure 2 Transport infrastructure). Additionally, we presented built-up areas (Figure 2 Buildings). The rest of the research area consisted of green areas, classified as nonpedestrian zones. The areas marked out constituted qualitative data. A map presenting the Then, we analyzed the research area in terms of transport infrastructure (Figure 2). Our priority was to mark out areas for pedestrians, i.e., sidewalks, routes for pedestrians and cyclists, parking lots and access roads. These elements were included in the research as areas in which pedestrian movement was allowed (Figure 2 Transport infrastructure). Additionally, we presented built-up areas (Figure 2 Buildings). The rest of the research area consisted of green areas, classified as nonpedestrian zones. The areas marked out constituted qualitative data. A map presenting the infrastructure of the

research area (Figure 2) was worked out with the application of the chorochromatic mapping method, traditionally used for presenting land use. infrastructure of the research area (Figure 2) was worked out with the application of the chorochromatic mapping method, traditionally used for presenting land use.

*ISPRS Int. J. Geo-Inf.* **2020**, *9*, x FOR PEER REVIEW 4 of 17

**Figure 2.** The map of the research area in terms of transport infrastructure: the chorochromatic **Figure 2.** The map of the research area in terms of transport infrastructure: the chorochromatic method (qualitative data).

#### method (qualitative data). **3. Methodology**

(X; Y).

**3. Methodology** To achieve our objectives, we adopted four main stages of research:


#### • Results: working out final maps (Section 3.5; Figures 8 and 9). *3.1. Concept of the Point-to-Polygon Transformation of Pedestrian Representation*

*3.1. Concept of the Point-to-Polygon Transformation of Pedestrian Representation* The employment of UAVs for recording the same area at different times makes it possible to scrutinize changes between individual images. Pedestrian movement is a highly dynamic phenomenon; hence, we have decided to use short time intervals between individual images of the research area. We adopted the interval of 10 seconds that would help us record pedestrian movement. The process of aerial triangulation of the images obtained, and their subsequent vectorization, allowed us to specify the location of individual pedestrians, which, in turn, made it possible to The employment of UAVs for recording the same area at different times makes it possible to scrutinize changes between individual images. Pedestrian movement is a highly dynamic phenomenon; hence, we have decided to use short time intervals between individual images of the research area. We adopted the interval of 10 seconds that would help us record pedestrian movement. The process of aerial triangulation of the images obtained, and their subsequent vectorization, allowed us to specify the location of individual pedestrians, which, in turn, made it possible to represent pedestrians by means of point objects and to identify each of them in the coordinate system (X; Y).

represent pedestrians by means of point objects and to identify each of them in the coordinate system

The ability to specify the accurate location of pedestrians in the images obtained and a large scale of the map resulted in the conclusion that the surface object would make the more accurate representation, allowing us to present the area occupied by an individual person in the way that resembles the actual state. To do that, it was necessary to transform the point representation of pedestrians into the polygon representation. To determine the area occupied by an individual

pedestrian, we have specified the value of a buffer by means of which transformation would take place. We assumed that the buffer value of a single point would be 37.5 cm (diameter of 75 cm), which corresponded with the largest length of the pedestrian's footstep (62.5–75 cm) [27]. Then, we marked out the area of the actual land use based on the obtained polygon objects representing pedestrians. The aim of this was to determine the areas informally used by pedestrians. With recording pedestrians at intervals, it is not possible to determine the location of individual pedestrians between the images obtained. In the research, we have assumed that obtaining images at short intervals and the average speed of pedestrians of 5 km/h [28–30] would allow us to determine the aggregate distance between pedestrians captured in individual images. Figure 3 depicts a concept of point-topolygon transformation of the pedestrian's representation. It shows the different stages of

**Figure 3.** The concept of point-to-polygon transformation of pedestrians' representation. **Figure 3.** The concept of point-to-polygon transformation of pedestrians' representation. *ISPRS Int. J. Geo-Inf.* **2020**, *9*, x FOR PEER REVIEW 7 of 17

**Figure 4.** Vectorization of pedestrians from the research area. **Figure 4.** Vectorization of pedestrians from the research area.

*3.4. Cartographic Visualization of Pedestrians' Location*

(qualitative data) (Figure 7).

of the infrastructure of the research area (qualitative data).

intermediary maps:

(Figure 6).

Mapping techniques were listed in terms of data type (geometry of objects).

In our attempt to meet research objectives on designing maps that use adequate mapping

• The map depicting land use with the location of pedestrians in the research area: the area method (qualitative data) and the dot method (quantitative data) (Figure 5). • The map depicting land use with the location of pedestrians outside the transport infrastructure: the area method (qualitative data) and the dot method (quantitative data)

• The map depicting land use with the areas occupied by pedestrians: the area method

Vectorization conducted in the research allowed one to obtain quantitative data presenting the location (X; Y) of pedestrians in the research area. The most natural way to present such data is to use dots [36]. For presenting quantitative data we used the dot method. Points as 0-dimensional objects are represented with a pair of coordinates [24], as presenting a point object on the map through direct physical representation of point geometry would make such object invisible. To read its location, it was necessary to use the additional attribute, the dot ratio. When designing our dot map, we assumed the dot width (diameter) of 0.5 mm, because it is the smallest symbol size recognizable by the map user [37]. The dot created is a cartographic representation of a point feature, and the geometric center of the dot corresponds to the coordinates of the point it represents. On the previously prepared map (Figure 5) pedestrians were represented as objects in the shape of a small circle, so that it is visible to map users. This is only the cartographic representation of a point object on the map. Quantitative data were presented in the background of the map showing the infrastructure of the research area (Figure 2). Such combination of methods of cartographic presentation made it possible to present both qualitative and quantitative data. As a result, researchers obtained a map for the creation of which two methods, the dot method and the chorochromatic map were used. In Figure 5 we demonstrated the spatial layout of all the recorded pedestrians (quantitative data) in the background

*ISPRS Int. J. Geo-Inf.* **2020**, *9*, x FOR PEER REVIEW 8 of 17

**Figure 5.** The map of land use with the location of pedestrians in the research area: the chorochromatic method (qualitative data) and the dot method (quantitative data). **Figure 5.** The map of land use with the location of pedestrians in the research area: the chorochromatic method (qualitative data) and the dot method (quantitative data). *ISPRS Int. J. Geo-Inf.* **2020**, *9*, x FOR PEER REVIEW 9 of 17

**Figure 6.** The map of land use with the location of pedestrians outside the transport infrastructure: the chorochromatic method (qualitative data) and the dot method (quantitative data). **Figure 6.** The map of land use with the location of pedestrians outside the transport infrastructure: the chorochromatic method (qualitative data) and the dot method (quantitative data).

incorrectly, i.e., informal areas. We specified that it was necessary to carry out the point-to-polygon

Transformation was understood as a change of the mapping technique, i.e., the transformation of the geometry of the element used in cartography, which would allow us to determine the area that pedestrians occupy outside the transport infrastructure specified. The average length of a human footstep is between 62.5 to 75 cm, depending on sex and type of walk [27]. We assumed that the area occupied by an individual moving pedestrian would be equal to a circle that is 75 cm (the maximum length of a human footstep) in diameter. That meant that each point object representing an individual pedestrian was been surrounded by a buffer of 37.5 cm, counting from the values of coordinates describing the location of the point. After transformation, one person was represented by means of a polygon object with the geometry of a circle occupying the area of 0.44 m<sup>2</sup> (Figure 7). The ArcMap *Buffer* tool was used to transform point features into area features. Transformation of the point representation of a pedestrian into the polygon representation resulted in the change in data type (from quantitative to qualitative data). The map presenting the area occupied by pedestrians outside the transport infrastructure (Figure 7) was worked out with the use of the chorochromatic mapping method. The buffer value adopted in the research constitutes just an example and can be modified,

transformation of pedestrians' representation.

depending on needs and the assumptions of the research.

A small research area and the UAV technology employed allowed us to obtain accurate location of individual pedestrians. We concluded that the representation of pedestrians by means of the dot (qualitative data).

*ISPRS Int. J. Geo-Inf.* **2020**, *9*, x FOR PEER REVIEW 10 of 17

**Figure 7.** The map of land use with the areas occupied by pedestrians: the chorochromatic method **Figure 7.** The map of land use with the areas occupied by pedestrians: the chorochromatic method (qualitative data). used the chorochromatic mapping method that presents qualitative data on land use. Additionally, aggregation allowed us to generalize areal data presenting the area occupied by individual pedestrians and to obtain the total area of informal land use (Figure 8).
