Organization and Correction of Spatial Data in Mobile GIS

Transcription

1 1514 JOURNAL OF NETWORKS, VOL. 8, NO. 7, JULY 2013 Organization and Correction of Spatial Data in Mobile GIS Feixiang Chen *, Xiao Ma, and Shaoliang Ni School of Information Science and Technology, Beijing Forestry University, Beijing, China Abstract With the development of wireless communication technology and Web GIS, the technology and applications of mobile GIS are also developed. Based on lightweight mobile equipment, mobile GIS has the feature of flexibility and convenience, and is suitable for field operations, but it is also low memory, low CPU, has limited display range and storage. Taking the particularity in mobile GIS into account, this paper presents a scheme for vector data organization and correction in mobile GIS. The paper consists of two parts: a organization form for vector data in mobile GIS, and a local real-time correction method of data based on the affine transformation. At last, the paper test the efficiency of the mixed form in organization and the accuracy of the correction scheme, and prove that both methods have certain feasibility. Index Terms Mobile GIS, Vector Data, Data Organization, Local Real-Time Correction I. INTRODUCTION Since the beginning of this century, computer technology and Internet communication technology have been rapidly developed, and the demand for information service is also increased. Applications of GIS gradually go into people's lives, and play an more important role in aspects such as field exploration, analysis of spatial data, car navigation, daily travel and so on. After the stage of C/S (Client/Server) mode and B/S(Browser/Server) mode, GIS is rapidly developing forward to M/S(mobile GIS) mode. Mobile GIS is a kind of Geography Information System running on a portable mobile terminal, the existing mobile device consists of laptop, mapping and acquisition instruments, car navigation equipment, handheld GPS, smart phone, tablet PC and so on, different devices vary in different aspects of performance, the combination of new technologies such as mobile terminal, GPS, wireless Internet and GIS greatly enriched the theory and technology of GIS, and expanded the applications. Compared with traditional GIS, the composition of mobile GIS are mobile terminal, wireless communication network, Geographic Application Server and spatial database. In terms of data processing, the storage and This work was supported by the Special Funds for the Scientific Research of the Forest Public Welfare Industry under Grant (No ). *Corresponding author, Feixiang Chen, bjcfxiang@163.com. computing speed of mobile GIS are limited, and the memory and CPU performance can not achieve the same indicators as PC. In order to avoid compatibility problem when exchanging data with PC and in order to avoid the result of data loss or error, we must design and use a more perfect organization and storage structure to carry out efficient and accurate data transfer. At the same time, in order to verify the quality and ensure the accuracy of passed data, it is necessary to adopt an correction method that is applicable and practical in vector spatial data, so that further inspections could be done. Currently, spatial data and attribute data are stored separately in most GIS software, namely mixed type data management. In the mixed type data management, spatial data is stored in the form of a binary, and that attribute data is managed by the relational database management system. Thinking of the storage capacity, computing power and screen restrictions of embedded devices, the software about GIS based on mobile device and PC are similar in basic functions, but there is a big difference in particular implementation, and a large part of this difference is reflected in the spatial data management. So this area is still explored at present. The organization of mixed proposed in this paper combines the advantages of several methods, so it could not only adapt to the special nature of mobile GIS, but also ensures data quality and efficiency. The proposed correction scheme also fully considered the characteristics of mobile devices, for its limited screen size and data volume, we proposed a local correction method to improve the speed of data correction. II. ORGANIZATION OF VECTOR DATA IN MOBILE GIS Vector data includes spatial data and attribute data, attribute data has a simple type and a clear format, usually use the relation database for storage; while the spatial data has a complex data structure, it takes up a large space, and is not easy to store and do reading operations. So this paper focuses on the organization of vector spatial data. Due to the low CPU speed, the small storage of SD card, the small display screen of mobile devices, and relatively the amount of data in GIS is large and need a large amount of computation, therefore the data organization model based on a desktop or workstation are short of completeness and adaptability for mobile terminal device[1]. So it is important to design a lightweight structure not only suitable for mobile doi: /jnw

2 JOURNAL OF NETWORKS, VOL. 8, NO. 7, JULY Header Accumulation s header Subfolders index Sub 1 Sub 2 Sub 3 Beginning of the accumulation End of the accumulation Figure 1. Accumulation schematic diagram Header header of composite page index (page1-2) Index of sub s (page 3) Page 4 Page 5 Page N beginning of compound end of compound Figure 2. Compound schematic diagram devices, but also could strongly meet with PC s compatibility, so that data could transfer between PC and mobile device without converting, and be operated efficiently. Here we will focus on the comparative analysis of several data organization model. A. Accumulation File The principle of accumulation is relatively simple, directly storing each sub- in accordance with the order to a parent, and the parent is called accumulation. The format of accumulation is shown in Figure 1. Accumulation s components include header, header of accumulation, index and sub s. File header: stored the version, format description and other basic information of this. Header of accumulation : pointed out the number of s that the accumulation contains and version information of the accumulation. File index: stored the starting offset of each sub- in the accumulation, the length of sub s and so on. From the above, this design is the simplest method to store multiple s into a single, when doing random reading and writing to the sub-, just need to do once offset calculating, so it has the highest efficiency, and the time complexity is maintained in O(1). But its drawback is that the location of each sub- is fixed, and cannot be extended, that is the sub s cannot be incrementally edited. To edit accumulation, each sub- must be first copied to a temporary folder as an independent, then to be edited, finally, when finishing editing, close the workspace, the sub s are re-merged, and wrote back to the accumulation. This method has little impact for the speed of the computer who has continuous disk read and write performance of at least 50MB /s or more, and helps protect data from being damaged by accidental editing or power failure. But for the mobile GIS platform which mainly uses flash memory chips as storage media, the speed of opening and closing workspace is unbearable, so accumulation can only be used as read-only format with high-speed, and must design another data format so that the can be edited directly on the mobile device. B. Compound File To make all information of a work area stored in a single, the way most in line with the logic of thinking and the most manageable way is the system in the, that is to further divide a into sub-s modeled on the principle of disk dividing. Generally we call the stored by divided sub-s as compound. Compound is the cornerstone of the OLE (Object Linking and Embedding) technology, software including Microsoft Office family and many other s formats are based on a composite. The interface of Windows composite document management is based on COM technology, but COM technology can not be applied on Android, and Microsoft did not disclose the detailed format of composite, so the platform should design and use its own compound engine, thus could guarantee the performance and maintain the portability to the most. The basic structure of compound is shown in Figure 2. File header: stored the version, format description and other basic information of this. Header of composite : basic information of the compound, such as page size, currently assigned pages and so on. Page index: indicates the page occupancy in the for distribution or release page. File index: stored the starting page number, the size, the allocated size of each sub-. Sub-s: sub are constituted by a series of continuous or discontinuous pages, starting page of each sub is the index page, which records the page number of the follow-up content to this document. By continuously allocate new page to different sub-s, the compound can be directly edited within the document. But it has the following problems: 1 Due to the characteristics of paging and sub- dividing, its reading and writing process requires a lot of analysis and calculation, so the efficiency is low especially when the s are trivial. 2 Frequent reading and writing will result in larger, the algorithm is complex, so it needs regular compression, and space release requires move all post content. C. Mixed File Due to the limitations in the performance of mobile platforms, accumulation is the most suitable format for its high efficiency, but the accumulation can not be edited directly within the document, so if modifying, addition or other operations are needed, it must first split accumulation into sub-s to be modified, then remerge them to accumulation after modifying. The essence of splitting is to copy the contents of the, when the size of the reaches to megabytes, the time delay of this strategy when opening and saving workspace is unacceptable, so accumulation strategy is commonly used for organization of s on PC.

3 1516 JOURNAL OF NETWORKS, VOL. 8, NO. 7, JULY 2013 TABLE I. DOCUMENT MANAGEMENT STRATEGY Behavior PC platform Mobile platform Open (Read-only) Directly open on read-only mode Directly open on read-only mode Open to edit Copy all sub-s to a temporary folder as a separate, and open from the temporary folder Directly open to edit Close (Read-only) Close directly Close directly Close after editing copy the independent contents of the temporary folder to save the accumulation Close directly TABLE II. DESIGN OF MIXED FILE FORMAT Header Header of mixed Section of Custom information size = ncustomdatasize Index of subfolders size = nfilenum* size of index Segments of accumulation Segments of compound Header of File checksum head, version and so on contains composite number of expansion or not sub-s chuselogicfile nfilenum Number of reserved subfolders index nfileindexspacecount Custom binary data segment Identification number Starting offset File size Length of custom message ncustomdatasize Starting offset of compound llogicoffset logical number of Expansion index in compound nfileindexlogicfileno Space of accumulation number of subfolders in compound Contents of a subfolder managed by compound module, the stored subfolders is the follow-up portion of accumulation s segment after incremental edit Compound can assigned pages flexibly, and facilitate the editing operations to sub- effectively, so it is a convenient and practical form of organization in mobile terminal. But data transmitted from PC to the mobile terminal is mainly formed in accumulate, if only the compound format is used in the mobile terminal, two different forms of organization would create compatibility problems. The conversion of data format not only reduces the efficiency, or even leads to data loss and error. Thinking of the format compatibility between PC and mobile platform, and in most cases we do showing and processing job instead of data editing on mobile platform, so we create an organization format which is the combination of accumulation and compound format. Add compound section to the end of accumulation, the compound is used to store the addition of each sub- in accumulation, thus could ensure the data reading speed from accumulation and make the editing and modifying job available within the. Strategy of document management in the two platforms is shown as Table Ⅰ. This strategy requires the docking of accumulation with the compound. In a mixed, sub-s can be divided into two parts, accumulation part and the compound part, and any part of the two parts may not be present. Here is a typical case: We make a pretreatment of a data on PC side, and save it as a.map, then copy it to an Android device, so that we could use the mobile GIS platform for field data collection and enrich the detailed data. For a.map has just been edited before copied to the mobile device, it only has the part of accumulation without the extension part of compound. When the field crews collect data, input data and save, the mobile platform needs to save the index, the coordinates of the location, the parameters and other information of the new added entity respectively to the indexer subfolder and the data subfolder of a map data, suppose that in a map data, the number of indexer subfolder is 1, the number of data subfolder is 2, so a compound must be attached to the back of a.map. Graphic data is saved in real-time, so data subfolder must create a new subfolder in the compound as its followup (called logical ), so we create a subfolder in compound and numbered it as 1. So far, the data subfolder of a.map is divided into 2 sections, the one is the no.2 in accumulation, and the other is the no.1 in compound, the two are used in combination. The indexer is similarly saved. Analyzing of the recorded data that the entire process needs to take, and thinking of the state of sub folders only belonging to accumulation or compound, an integrated design of the format of mixed is shown as Table II. In order to avoid running out of the number of subfolders, a logical could be specified in the compound through nfileindexlogicfileno to continuously store the index. In order to make some common information be read directly instead of getting information from the analysis of a "mixed processing module, Custom binary data segment is set. The starting position of the segment is fixed, so it could be read directly without parsing. Each sub- is stored in the accumulation segment, if they were opened on a mobile device, the accumulation segments will be immutable. So if it still needs to write data to a subfolder, a logical subfolder should be created in the area of the compound segments as extension. This extension part is created only when needed, the sequent number of this logical

4 JOURNAL OF NETWORKS, VOL. 8, NO. 7, JULY subfolder is stored in number of subfolders in compound of compound s subfolder index. According to the design, all s saved from PC side are accumulation s, while editing in the mobile platform, the increased portion of the is a compound. This mixed model mostly ensured the compatibility and read-write efficiency of spatial data in the mobile platform. III. DATA CORRECTION IN MOBILE GIS Mobile devices have the characteristics such as easy to carry, suitable for a wide range of work environments especially the field work, which make it valuable for field geological survey work, and can reduce the burden of field work and improve the efficiency of data collection. The field geological survey needs to check if the digital maps match with the actual location and geological environment, to ensure the correct basis of the data collection and survey, so correction of spatial data on the mobile terminal is required, and we should compare and contrast the coordinates in existing map with GPS measured position. While, contrast to the large spatial data in geological field work, there is much insufficiency like limited storage space and low computing speed in mobile terminal, so it is difficult and has low feasibility to do the map correction in accordance with the general process of spatial data correction in the intelligent mobile terminal, which is a huge amount of calculation task. A local correction scheme has been designed in this paper, it only corrects the portion displayed in the screen, thus reduces the computation greatly, and achieves the real-time correction of spatial data in mobile terminal and the GPS points. The calibration process is essentially a process to find a mathematical relationship (or a function), that is the graphic coordinates conversion before and after transmission. Suppose that graphic coordinate before converting is (x, y), and the graphic coordinate after converting is (x, y ), then the mathematical relationship is generally described as following formula: x' a0 a1 x a2 y a11 x a12 xy a22 y A (1) y ' b0 b1 x b2 y b11 x b12 xy b22 y B This mathematical relationship is often represented as a binary polynomial in once, twice, three times, and higher-order expression. Which one to select in actual calibration processes depends on the specific deformation method: if it is a linear relationship such as translation or scaling, it is better to use affine transformation (linear translation), the formula is a special form of multiple expression as formula 2; or if it is quadratic or higher form, it would be better to use quadratic transformation as formula (3) or the transformation of high-order. x' a1x a2y A y ' b1x b2y B (2) x' a a x a y a x a xy a y y ' b0 b1 x b2 y b11 x b12 xy b22 y Besides, the error may be caused by stretching, rotating or other artificial or incidental operation in the process of digitizing, and they will cause errors inconsistent with the actual such as changes in topological relations, changes in node consistency, and changes in entity uniqueness [2]. For instance, in order to solve the error caused by rotational transformation, set XOY as the coordinate system before correction, and set xoy as the coordinate system after rotation correction, then the formula of coordinate transformation is shown as formula 3. x X cos Y sin y Y sin X cos Alpha in the above formula can be obtained through the common points coordinate of the two systems. The expression is shown as follows. (3) (4) tan ( yx xy ) / ( xx yy) (5) The correction process designed for the mobile terminal: Step 1: Select the control point. Select a certain number of GPS positioning points and the corresponding points of the map in mobile platform as the control points. Stet 2: Converting the coordinates of GPS positioning to make its coordinate type as same as the coordinate type of spatial data in mobile map. What GPS get are the latitude, longitude and elevation values, so it is supposed to carry out projection calculation and other operations to the data. Step 3: Establish affine transformation between GPS positioning point and spatial data point. On the basis of GPS points, create the affine transformation from a point of the spatial data to the GPS positioning point using the formula 2, where the coordinates (x, y) is a point of the spatial data, and coordinate (x, y ) is the corresponding GPS point after the conversion. Step 4: Use least-squares fitting to find the parameters a1, a2, A, b1, b2, B, and then get the transformation model from the spatial data to the GPS map. Suppose that the transformation is called P Trans, and then spatial data can be corrected through P Trans based on GPS positioning point. Figure 3. Schematic diagram of P trans Step 5: Seeking the inverse transform of P Trans as P Trans, so that the corrected spatial data can be restored as the original spatial data before correction through the inverse transform P. The formula is shown as follow:

5 1518 JOURNAL OF NETWORKS, VOL. 8, NO. 7, JULY 2013 x a1' x' a2' y ' A' y b1' x' b2' y ' B' (6) by the mixed, three sets of spatial data sized of 1M, 2M, 5M were selected respectively to do the test. When they were organized respectively by the accumulation, the compound and the mixed, the reading time of each kind of structure is shown as follow: TABLE III. SPEED COMPARISON OF DIFFERENT FILE FORMATS Figure 4. Schematic diagram of P trans Step 6: After correction through P Trans, the corrected spatial data will be displayed on the screen. Use P Trans on the displayed part in the screen to push back and obtain the original area, and we just need to do correction on this part of original spatial data, can we get the corrected data and display it on the current screen, thus we could also achieve a desired effect. Figure 5. Multi-level incremental storage Figure 6. Effect before and after P Trans Step 7: After calculating out the area to be corrected, solve the minimum bounding rectangle of this part, and then correct the original spatial data within the rectangle, that is, only to correct the data within the rectangle through P Trans, can meet the display requirements. In summary, according to the area to be displayed on the screen, do P Trans for the corrected map to obtain the corresponding original image, then convert the part of original image that is within the minimum bounding rectangle with P Trans instead of correcting the whole spatial data. The local real-time correction in this paper can effectively reduce the pressure of mobile device, and effectively increase the computational efficiency to the data. IV. TESTS OF ANALYSIS AND VERIFICATION A. Efficiency Test of Mixed File The advantage of a data organization form is reflected by the reaction efficiency and the reading speed of the organized in this form. In order to verify the efficiency advantages of spatial vector data that is formed Data size Accumulation Compound Mixed 1M 0.48s 0.39s 0.42s 2M 1.07s 0.81s 0.84s 5M 2.59s 2.10s 2.15s In the case of a small amount of data, the memory occupied by data processing and the storage space are relatively small, so the comparative results in PC could also reflect the situation in the mobile device. It can be seen from the above table that mixed is significantly better than the other two data structure on the reading efficiency. Furthermore, in order to fully demonstrate the superiority of mixed in the practical application of mobile GIS, we use the mobile device based on Android 4.1 system to test the data open time and data edit time respectively in GeoGIS and MapGIS, the organization of spatial data is formed as mixed in GeoGIS. The efficiency comparison of the two is shown on the follow table: Time consuming GeoGIS MapGIS 1M 0.42s 0.56s 2M 0.84s 1.03s 5M 2.15s 2.47s As can be seen from the experiment, the mixed has a higher efficiency than any other form of data organization in mobile GIS. B. The Test of Correction Accuracy The distance error before and after data correction can be used to test the value of the data correction. In mobile GIS, the distance error between the map location that is imported into mobile terminal and the actual position data obtained by GPS measurement could be judged by the distance of landmarks. Select a specific landmark like a salient feature or a cross point, then read its position in the map as coordinate A, and measure the actual GPS coordinate of this position as B, calculate the distances between A and B respectively before and after correction. Assuming that the coordinates of the two points in the same position are A (x1, y1), and B (x2, y2), then use the Euclidean distance formula to calculate the distance, it is shown as following formula: AB x x y y ( 1 2) ( 1 2) (7) Here we choose four points A, B, C, D on the following table as mark points, where, the "GPS measured coordinates" indicates the coordinates transmitted from the GPS measured values by Gauss projection, and it is measured by meters. So the coordinates of the four points and the distance error before and after correction are compared in the following table 4:

6 JOURNAL OF NETWORKS, VOL. 8, NO. 7, JULY TABLE IV. DISTANCE COMPARISON BEFORE AND AFTER CORRECTION Coordinate before correction GPS measured coordinates Coordinate after correction distance before correction Coordinate after correction A (492168, ) (492181, ) (492180, ) B (493426, ) (493412, ) (493411, ) C (493125, ) (493119, ) (493120, ) D (492761, ) (492777, ) (492779, ) It can be seen from the above table that after correction, the distance between the spatial data and the actual coordinate is greatly reduced. That is to say, the correction method of spatial data proposed in this paper plays an obvious role for data correction in mobile GIS. V. CONCLUSION With the development of geographic information system, Mobile GIS reflects a growing value in people s life and work because of its convenience and practicality. While Mobile GIS has the particularity that is different from the traditional geographic information systems, for instance, the hardware and software configuration of mobile terminal make it just adapt to run low memory and low CPU environment, besides, the limited storage space in platform itself makes a sharp contrast to the huge volume of spatial data. In this paper, an organization form of mixed is adopted, it effectively solves the problem of data storage and read-write efficiency of mobile terminal, and we adopted a method of local real-time correction to do efficient and accurate correction for vector spatial data, so could facilitate field operations. Experiments show that both methods can achieve the efficiency and accuracy that mobile GIS operations required for the application, so it has the certain practicality. REFERENCES [1] L. Q. Dai, N. Chen, J. Q. Yang, I see the future of GIS, Geospatial Information, vol. 9, no. 1, [2] X. Q. Zuo, Y. C. Zhang, C. Feng, A compute method of road travel speed based on mobile phone handover location, Journal of Networks, vol. 7, no. 10, [3] J. H. Zheng, L. Yan, Y. F. Liu, L. N. He, Study on data organization of personal navigation services, Computer Engineering, vol. 32, no. 24, [4] H. X. Liu, H. D. Liu, H. J. He, Y. L. Zhao, Development and application of mobile GIS, Cartography and Geographic Information Science, vol. 3, no. 4, [5] R. M. Tang, The study on afforestation operation design system based on PDA, Master Thesis of Beijing Forestry University, [6] X. L. Shi, The research and design of mobile geographic information system based on PDA, Master Thesis of Sichuan University, [7] G. Tian, X. H. Tong, J. Zhang, Investigation of organization model for Vector Data Based on PDA, Journal of Taiyuan University of Technology, vol. 35, no. 1, [8] D. R. Li, J. Y. Gong, F. L. Bian, Data organization and processing methods of GIS, Bulletin of Surveying and Mapping, no. 1, [9] Y. Zhao, H. Chen, S. G. Wang, An improved method of detecting edge direction for spatial error concealment, Journal of Multimedia, vol. 7, no. 3, [10] G. M. Kuper, J. W. Su, A representation independent language for planar spatial databases with Euclidean distance, Journal of Computer and System Sciences, vol. 73, pp , [11] J. H. Guo, Z. W. Yao, Y. L. Bao, W. S. Zhang, An automatic correction algorithm of vector map, Chinese Journal of Image and Graphics, vol. 4, no. 5, [12] B. Z. Wei, Z. M. Zhao, X. Peng, Spatial information based medical image registration using mutual information, Journal of Multimedia, vol. 6, no. 3, [13] D. P. Xu, S. S. Zhou, T. Zhong, Correction algorithm of vector map based on matrix transformation, Journal of Wuhan University of Technology (Transportation Science & Engineering), vol. 28, no. 2, [14] J. Zhang, Research of several key techologies of changeonly information collection system on field based on PDA, Master Thesis of Central South University, [15] Q. H. Su, H. Cheng, W. B. Sun, Overview of remote sensing image registration methods, Infrared, vol. 31, no. 10, [16] P. Wang, Mobile GIS and its new applications, Geomatics & Spatial Information Technology, vol. 29, no. 5, [17] X. Wang, W. Li, Discussion on the fuzzy comprehensive evaluation method of the spatial data quality, Modern Surveying and Mapping, vol. 34, no. 3, [18] J. L. Zhao, Z. K. Feng, Research on error correction and precision analysis of map scan vector spatial data, Mine Surveying, no. 4, [19] W. C. Fang, Z. J. Zhang, R. Zhang, Y. Q. Luo, X. C. Hong, Z. J. Shao, A combination method of least squares for data correction, Computers and Applied Chemistry, vol. 29, no. 7, [20] S. Liu, J. X. Pu, L. Zhao, Localmotion correction for functionalmagnetic resonance images, Journal of Computer Applications, vol. 29, no. 11, Feixiang Chen was born in Hubei Province of China in He received the M.S. degree in computer applied technology from China University of Geosciences in 2003, and received the Ph.D degree in cartography and geographic information system from Chinese Academy of Sciences in Now he is an associate professor in Beijing Forestry University, and his main research field is mobile GIS. Xiao Ma was born in the city of Taiyuan, Shanxi Province, China, in She is going to graduate and receive her bachelor degree in management from the information management and information system major in Beijing Forestry University in Currently she is a four-year undergraduate of Beijing Forestry University.

7 1520 JOURNAL OF NETWORKS, VOL. 8, NO. 7, JULY 2013 Shaoliang Ni was born in Tahe, Heilongjiang Province, China, in He received his B.S. degree in management from the information management and information system major from Beijing Forestry University, Beijing, China, in At present, He is studying for his master's degree at Beijing Forestry University.