To say the data source, first of all to do a classification of map data, because the data of different classification, its source, collection methods are greatly different.
I don’t want to say that @Liu Le’s classification method is wrong or not possible, but that its classification method may not be too appropriate and reasonable to fully explain this problem. There are also some minor problems with some of the ideas and descriptions in the book, so I will make some corrections and more targeted supplements to the questions, hoping that you will not be misled by some false concepts.
To understand the classification of map data, one must first understand the concept of map layers:
Review images
As shown in the figure above, the expression of the electronic map to our actual space is actually described by different layers, and then expressed through the overlay display of layers.
For each target we apply to the map, the overlay layer is also different to show the information we need for the target.
Secondly, I introduce the concepts of vector model and grid model. GIS (electronic map) uses two different mathematical models to simulate the real world:
Vector model: with multiple X,Y (or X,Y,Z) coordinates, the natural objects through points, lines, planes to express
Review images
Grid model (tile model) : Use squares to simulate entities
Review images
We are now in the Internet public service, or see, in the vast majority of mobile phone APP is based on the grid (tiles) model mapping services, such as baidu or Google map, we see in fact description of a certain area, by 10 layers is 20 layers of images of different resolution, when a user to zoom in, According to the zoom level, select tiles of different resolutions to join together into a complete map (due to the general public service, tiles are downloaded from the server, when the Internet speed is slow, users can actually see the switching and stitching process of pictures of different resolutions)
For vector-based electronic maps, as all the data are stored and managed in a vector way, in fact layers are a relatively weak concept, because any map elements and data can be freely classified or divided into different layers according to needs. The relationship between the layers can be complex, for example, all the road data can be made into one layer, or the main road can be made into one layer, and the branch road can be made into another layer. Data can be grouped and combined freely in layers.
For raster models, the concept of layers is very important. Since layers are generated and the elements contained within each layer are relatively fixed, the concept of a base image is introduced. That is, it is a layer that contains the most basic and commonly used map data elements, such as roads, rivers, Bridges, green Spaces, and even some base maps that contain Outlines of buildings and other features. On the basis of the base image, you can overlay all kinds of layers we need to meet the needs of the application, such as: traffic jam layer, satellite image, POI layer and so on.
The base map is usually rendered by selecting the necessary map vector data items and then through the work of the map artist, setting the colors, fonts, display methods, display rules, etc. (usually rendering a whole set of tile maps with different resolutions)
Of course, even in tile services, simple vector layers can be overlaid on top of the tile base, such as road directions (essential for navigation and route planning), and POI points (find a restaurant and gas station, etc.). However, tile engine cannot construct all map data in the same spatial data engine, so it is difficult to carry out complex map analysis and processing.
So with all the limitations and drawbacks of tile engines, why not just use vector engines? Because tile engines have significant advantages:
1. Can load large-scale concurrent users, vector engine to consume a lot of server computing resources (because there is a complete spatial data engine), even if only dozens or hundreds of concurrent users, need extremely exaggerated server computing capacity. Vector engines cannot meet the requirements of public Internet services. 2. Due to the rendering work of map artists, tile maps can be made very beautiful and easy to read, which is more suitable for ordinary users to browse
Attached: a vector map screenshot:
Review images
Ok, so much said, the main purpose is to introduce the concept of layers and base images to facilitate the classification of map data below
In order to explain data sources, collection channels and collection methods, I divide map data into the following types:
1. Base map data: in fact, it is the most basic ground object shape data and certain related additional information (such as road name, river name, etc.) in the map. As a matter of fact, with the progress of remote sensing and aerial photography technology, the proportion of such data relying on field collection has become smaller and smaller. Commercial map data providers, especially Autonavi, have rarely adopted field collection in consideration of cost and benefits. The data in this part mainly come from three types:
Official map: Strictly speaking, this can not be said to be a separate channel, because the data of official map itself also comes from the following two channels, but official map generally comes from the authoritative mapping and release of relevant government departments, so it is also considered as a single channel. It should be noted, of course, that the maps that map manufacturers can obtain or buy from state authorities are much more detailed and varied than the ones we buy in street shops, often in electronic format. Of course, truly accurate maps (e.g. 1:200 scale or higher) are restricted from being published in any country. (For reference, China’s regulations on the Internet can be published on the map, the highest accuracy is 1:10,000)
Review images
On-site mining: it is the use of professional instruments and meters in the field environment mapping obtained. Such a collection method is very time-consuming and time-consuming. First, it costs a lot of money, second, it has a long cycle, and third, it requires a high environment (go to the Himalayas to measure blood), and it may not be able to completely keep up with the urban changes in China. But the advantage is high accuracy, high confidence, very high accuracy. This is the main means used by national surveying and mapping departments, for a city like Beijing, it is generally a few years to complete a round of surveying and mapping. Generally for most commercial mapping, only in a small number of local needs, the key mapping can be used.
View the picture this you should occasionally see on the road
Of course, in places where the accuracy and accuracy requirements are not so high, field collection can also use some cheaper and more convenient tools, rather than professional mapping equipment. Such as walking with a handheld smart device that carries a high-precision GPS or other positioning to draw Outlines.
Aerial health film production: it is through the high precision aerial photos or satellite photos or remote sensing photos taken or purchased by oneself, on the basis of this as a negative for artificial vector marking and outline, so as to form their own vector data. Now the precision of aerial film or remote sensing film can be very high, generally speaking, it is easy to achieve the accuracy of 0.05 meters. Autonavi’s own aerial photographs are said to have achieved accuracy of 0.03m, which is often sufficient for commercial map data. Even as the national authority of surveying and mapping, surveying and mapping in a large number of wilderness mountains, also mainly rely on this means. At present, the commonly used aerial or guard photography means include airborne digital camera, airborne remote sensing and 3D laser scanning (mainly used for 3D map data acquisition).
Review images
0.05m precision aerial film
Review images
Sanitation road network labeling
Review images
Review images
Aerial/photographic annotation and sketch, the first is the operation on the negative, the second is the vector map after the annotation
Review images
Schematic diagram of data processing and production (from a public data of Autonavi)
From this part of the data, Baidu is not their own collection production capacity, there is no license (no mapping qualification). This piece of data of Baidu is basically to four dimensional graph new buy.
For domestic data, there are mainly two suppliers, namely Autonavi and Siwei Tuxin.
The relationship between Four-dimensional tuxin and the national surveying and mapping unit is unusual. Its data depend on the national surveying and mapping unit to supply the majority of the data (of course, there are also mutual supply).
Autonavi also has some data from national surveying and mapping units. However, autonavi has good capability in aerial photography production (it has also undertaken surveying and mapping tasks of some national surveying and mapping institutions). Comparatively speaking, the proportion of relying on data from national surveying and mapping units is lower.
Generally speaking, the collection and production of this part of data requires the qualification recognized by the state in China. There are not many commercial institutions with the qualification except the national surveying and mapping agency, and the two companies that are really active in the data supply market are autonavi and Siwei Tuxing.
Other data, whether Google Maps or Apple Maps, are basically purchased from the two companies.
2. POI data: strictly speaking, it belongs to vector data, but it is the simplest vector data. In other words, it is coordinate point annotation data. It is also the most commonly used data layer on electronic maps.
The data we use on electronic maps on a daily basis is POI data (a balloon dot on a map).
POI data is only the data of coordinate points associated with information, which does not involve lines and planes. POI data is the simplest vector data, which is used for simple site annotation without corresponding requirement of ground object contour.
The content of POI data is varied. Generally, POI data provided by POI data suppliers are commonly used daily site data, such as restaurants, shops, gas stations, banks and other commonly used daily facilities.
Of course, in some special map application areas, you can also entrust these data providers or collect POI data for special purposes, such as manhole covers, fire hydrants, etc
Tax GIS system annotates enterprise and tax information
It is worth pointing out that POI data is easy to edit and update, and is often used for dynamic data annotation, the most classic of which is vehicle positioning annotation.
POI data are collected and produced from various sources, which cannot be described completely. In general, there are mainly the following:
- A) Sweep the street for continuous shooting by integrating GPS cameras on foot or by car, and then manually input and mark according to the shooting results after going back. This method is suitable for large-scale collection and mark, with high efficiency and low cost and the majority of car shops, especially for the collection and mark of storefronts and places along the street. It is one of the main acquisition methods of data acquisition suppliers
- B) Take photos (mainly for forensics), input, submit and collect by full-time or part-time personnel using handheld smart devices with GPS (such as smart phones). This collection method is mostly used as a supplement to the above method A. Use in areas inaccessible to vehicles or in areas where business facilities change frequently
- C) Address reverse compilation: use the algorithm to mark the location through the address number of the house plate and the road data in the vector map. The labeling accuracy is relatively low and the accuracy is not high, but the cost is very low. It is used in the acquisition field which does not need special high precision and cost control is strict. Everyone in the map service search box input address number, you can directly appear marked point, with this technology.
- D) Internet or enterprise acquisition: directly capture or purchase from some professional service websites (such as Dianping and Ctrip), or directly screen and obtain from people’s marks on their open map services. This is what Google, Baidu and Autonavi themselves can gain by opening up their mapping services to the public for free. This acquisition is especially valuable for use by free enterprise customers with open apis.
Domestic POI data suppliers do not have too many qualification restrictions, compared with the base map data suppliers, there are many more, such as Tuba are POI data suppliers, of course, SiWEI Tuosun and Autonavi also provide POI data, each POI data supplier has its own classification method, data definition and other content. In many cases, we also buy from each other and complement each other.
Baidu map data in this respect, mainly from four-dimensional map new and daotong, of course, there are other sources, and even a small amount of self-produced data.
Amap mainly produces its own data, supplemented by purchasing from some professional service providers (Koubei, Dianping, Ctrip, Letu, Soufun).
3. Other data layers or data: the common ones are the map layer, traffic condition layer, 3D map and street view map. Some professional areas are building maps, indoor maps, temperature distribution maps, business district distribution maps, topographic maps, hydrological maps and so on.
Review images
The micro map
Review images
Real estate market map
Review images
Address hazard map
Review images
TV Limited network distribution management chart
Review images
Indoor figure
Review images
Statistical report chart
Review images
Population density map
Review images
Three-dimensional real map
Review images
3d data schematic diagram
Are so many multifarious figure, mainly is to illustrate, based on the electronic map data layer is a wide application scope and application fields, different layer, represents the different data, this area has a large number of professional application and data, the acquisition method, sources is multifarious also, hard to do.
A few common sources of data:
- A) Traffic congestion data: this data is generally obtained from professional data suppliers, which have deep cooperation with traffic departments. Their data collection mainly relies on GPS installed on taxis to collect real-time speed, or speed measurement through cameras, infrared probes and radar is supplemented
- B) 3D data: mainly rely on laser scanning and manual modeling processing
- C) Fake 3D data (the kind of 45-degree 3d top view that cannot be rotated) : manual production depends on photo shooting and material stitching.
- D) Street scene: rely on spot shooting
Baidu Map basically only has the most basic and common layer data, such as part of 3d data, traffic layer data, health layer data, etc. Baidu does not have the collection and production capacity of this part of the data, are purchased from different suppliers.
Amap has some special data, such as real estate data, etc. Autonavi has collection and production capacity in some data fields (such as THREE-DIMENSIONAL data, etc.), and can carry out professional collection and production process and provide professional layer data according to customer requirements.
However, some special partial, special professional data fields (such as the above have television network, address disaster map) need to use their own professional tools to make
There are many special data layers that cannot be displayed on tile engine, or that don’t make sense on tile engine.
It only exists in vector engine applications.
Conclusion:
Baidu’s map data is mainly bought, while Autonavi’s data is mainly collected and produced by itself.
As far as the domestic situation is concerned, the main data are collected. This point and foreign developed countries have quite big difference. In foreign developed countries, because the construction speed is relatively slow, the level of government information and information transparency is better, in fact, there is no need to collect so much work.
There is a saying in this industry, called foreign (developed countries) to internal work, supplemented by field, domestic to field work, supplemented by internal work.
Due to the relatively slow change of land features, the data published and released by the government is timely, transparent, accurate and usable. Therefore, many data manufacturers in this industry abroad can directly process the data published by the government, and there are not many changes. In China, they still rely on data collection by themselves in data production. (for example, it is typical of acme is Japan, where the new installed a traffic light will be on the government web site reported accurately and in time, and marked position, not to mention building construction, urban construction and other big moves, so the data producer take these data can be directly used, rarely have to collected)
By: Tianguang via: Zhihu Daily
End.
Baidu Map, Gaode Map data from where to get?