The Data Capture Challenge - Innovative Solutions
Philip Pickford New Zealand Aerial Mapping Limited PO BOX322 Albany Auckland New Zealand Abstract This paper focuses on innovative in-field data capture techniques that have solved real problems facing utilities. These techniques have been developed and refined by way of experience gained in a number of large scale data capture projects. Utilities considering data capture projects face numerous challenges, including scope, methodology, cost and quality control. These challenges are discussed, along with options available which conform to a number of practical and user-defined principles. Global positioning systems (GPS), gee-referenced imagery and data entry are discussed, as they pertain to in-field capture, with an emphasis on using this technology in different and innovative ways. The paper uses a case study to expand on the principles and options discussed. The project cited is the largest data capture project undertaken in New Zealand and involves positioning and attribute capture of over 100,000 power poles. This project has been especially challenging because of factors such as difficult and remote terrain, tight budgetary constraints and very challenging timelines. Data capture for this particular project is discussed, covering methodology, digital topographic landbase use, digital gee-referenced imagery capture and quality control procedures. Principles of Data Capture Data capture is best defined as the means by which information relating to physical entities (the data) is transferred into a digital format. This transfer can either originate from plans where the information contained in the plans is converted directly into a digital form with no additions (data conversion), or where the plans are used simply as a guide to in-field data capture where the information is gathered as a result of direct observation of the particular physical entity (field inventory). Invariably, data conversion involves the digitizing or scanning of existing plans producing a digital copy of those plans. Using this method there is little opportunity to improve or add to the data. Any errors, inaccuracies or shortcomings in the existing plan data will be reflected in the converted digital data. Using this method of data capture is therefore very dependent on the quality and completeness of the existing plan information. Field inventory involves visiting the plant on-site and using that visit to observe the attributes that need to be captured. Existing plans can be used as a guide, however, the actual data captured is directly related to on-site observations. This data needs to be input on-site as well, ensuring real-time observation and input. This method allows for extensive attribution; anything that can be observed and confirmed can be captured. It also gives the opportunity for considerable data improvement, both in terms of accuracy and completeness. This paper focuses on field inventory as a practical and cost effective data capture methodology. When discussing data capture principles, two perspectives must be reviewed: the perspective of the organisation that will ultimately use the data (The Client), and the organisation that will physically capture the data (The Contractor). For a successfid data capture project these two perspectives must align and be complementary, thus allowing for clienticontractor partnerships to develop rather than the more traditional situation of argument and conflict. The Client The foremost principle fi-om the client’s perspective is to ensure that clear and well defined objectives are in place that will ensure the key business drivers for automated mapping, facilities management, geographic information systems (AM/FM./GIS) within the organisation are satisfied. These objectives will in turn dictate the information required, the appropriate methodology and quality standards of the data captured and the acceptable cost parameters. The Contractor Of prime importance to a contractor is to ensure he completes the data capture contract to the satisfaction of the client. In order to facilitate this result he must ensure he is fully aware of the clients specific requirements and the data capture methodology he uses must be able to meet those specific requirements within the cost constraints imposed by the client. The response to this simple challenge is the difference between a successfid project and a failed project, and can best be enlightened by fostering a close and productive client/contractor relationship. The contractor must work within his own constraints to ensure successful completion. These fundamental constraints are the product of experience and detailed analysis: Single Site Visit Visit the site once to capture data. Ensure that all data is captured and all tasks are completed at that single visit. In short - “the do it once, do it right principle”. Data Entry Methods of data entry must ensure a balance of three factors - speed, accuracy and cost. Data entry must be on-site to ensure the closest correlation between on-site observation and data entered. Digital Landbases An integral part of field inventory is the use of digital landbases. The use of a photogrammetrically derived digital landbase of physical features with a defined and absolute spatial accuracy level allows on-site confidence of real-time GPS with the opportunity of absolute gee-referencing of plant. As a result all positional quality assurance is real-time and on-site. Quality Assurance On-site quality assurance must be maximised, allowing the principle of “do it once, do it right” to be confirmed. However, a method of off-site quality assurance must also be possible. This area has been of considerable challenge but technology has been able to provide solutions, primarily in the area of gee-referenced digital imagery. Case Study - The Client’s Perspective Powernet Ltd is a joint venture company which has been formed to manage, but not to own, the electricity networks in the province of Southland. The province spans 28,000 square kilometres (10,800 square miles) over the lower end of the South Island of New Zealand. The merger was in response to the reform and deregulation of the electricity distribution industry in New Zealand and is only one part of a total restructure of the incumbent electricity companies in Southland. The shareholders and directors of Powernet recognised that the deregulation of the electricity industry would lead to a rationalisation of the number of power companies in New Zealand. This would seem inescapable when you consider that New Zealand has 44 power companies for 3 million people compared to England which has 12 power companies for 60 million people. In order to manage the network efficiently, Powernet required accurate and comprehensive information on network plant and equipment, including an assessment of its present condition and an increased knowledge of the network customers’ requirements. This information had to be accessible to all parts of the organisation and be reliable to the extent that it could be used as the basis for better informed business decisions. Powernet had inherited paper-based information systems that were deteriorating, time consuming to maintain and inaccurate. It became obvious that these systems required major rationalisation to allow the company to meet its fiture information needs. Powernet therefore decided to embark on a project that would provide a network information management solution through the employment of AM/FM/GIS technologies. The project was thus coined the Powernet Information Management Solution or PIMS. The primary objective of the project was: “to significantly enhance the content, accuracy and currency of network asset data, and provide universal access of this information to Powernet and its associated contractors”. The key business drivers for the project were:
Project Phase The PIMS project commenced in February 1995 with an aggressive implementation programme consisting of 4 distinct phases over a 3 year period. Phase one began with a pilot project which involved the capture of data from an area approximately 500 square kilometres (200 square miles) containing 7000 poles and 5500 customers. This was a critical phase in the project as it provided an opportunity to confirm the key benefits to the business and test processes prior to full implementation. Phase two involves data capture of the entire network, consisting of approximately 100,000 poles and 8,500 distribution transformers. This phase encompasses an area approximately 20,000 square kilometres (8,000 square miles) and represents the largest geographical project of this type ever undertaken by a Power Company in New Zealand. Part of phase 2 is to extend the PIMS facilities out to a contracted network maintenance work force predominately for in-vehicle use. All 15 contracting companies will be users by the first quarter of 1997. Phase three will involve the development and integration of fiu-ther sofltware applications to meet Powernet’s secondary functionality requirements. Phase four will ensure the extension of phase 3 functionality to all users across Southland. The PIMS project team spent considerable time defining the data requirements necessary to meet Powernet’s diverse needs. The data required varied in nature from plant condition based data, connectivity data, pole attribution data, positional data and physical landbase features. Potentially as many as 80 variables can be gathered from any one pole site. Having specified the data requirements it became apparent that the existing maps for above ground plant were not suitable. They lacked detail, were inaccurate and not current. Field inventory was therefore the data capture method of choice. The existing plans for the underground network, however, were very well detailed and met the required accuracy standard. Consequently these were digitised. Key Mimaganmt Processe s Powernet recognised at the commencement of the project that the key critical success factor was not the hardware or the software solution that was chosen, but how the project was managed. For this reason Powernet was determined to manage this project efficiently over a relatively short implementation time frame and put in place the following processes and procedures to achieve this end:
As previously stated, the prime motivator for the contractor is to ensure completion to the satisfaction of the client. Hence the client’s specific requirements and the data capture methodology that is able to meet those requirements are paramount to successful completion. The PIMS Data Capture Services Contract was very specific and quantitative in the two important areas of scope and data acceptance criteria. These requirements provided the challenges specific for this project and in turn provided the catalyst for innovation. Specific challenges included aggressive timelines, some difficult and remote terrain and tight budgetary constraints. The areas of innovation that addressed these challenges were proprietary data capture software, the use in the field of physical digital landbases, portability of equipment and quality control methods that highlighted the effectiveness of digital imagery. Data Capture Software The aggressive t’ming of the project requires 100,000 poles to be positioned, labeled I and attributed in 12 months. Consequently, data capture rates needed to be predictable and attainable. The development of proprietary data capture software allowed for advances in both speed and accuracy. This software development included input from Powernet with respect to attribute structure. The key features of this software are listed. User Friendly The package needed to be simple and user friendly thus allowing for effective results from personnel with limited experience. This factor of personnel experience was an important issue in the early stages of the project. Initially some experienced operators were available from other projects, however, the majority of operators had no industry experience and limited computer experience. This was a consequence of the available employment pool rather than a planned strategy. The data capture software package had to therefore be based on providing pick lists and options rather than complicated and time consuming data entry. Geographic Link The software provided a link between the data captured and the relevant spatial information. As a result operators are able to view spatial relationships on-site using real-time GPS and physical landbases, while capturing other attribute information. 297.This fiction has been of importance in on-site quality assurance of positional accuracy. Quality Assurance The package is also capable of informing the operator of logic errors and asking for confirmation of unusual plant combinations. It also allowed for the capture of digital images of pole structures, gee-referenced to the poles’ coordinates. This allowed for later off-site quality assurance of important and visible features. Flexibility The attribution framework within the sofisvare is flexible and able to be client specific. Changes to this fiarnework in the early stages of the pilot project were able to be performed rapidly and on-site. Digital Landbases A digital landbase of physical features, at a defined accuracy, is the chassis by which any successfid AM/FM/GIS system operates and as such must be an integral part of the data capture process as well. It gives the data spatial integrity. If that integrity is available at the time of capture, quality assurance advantages are substantial. It allows for real-time positioning checks, checks of positional relativity between items of plant and orientation of items like pole guys and street lights. The landbase used in the PIMS project differed between rural and urban areas. In urban areas, kerblines, property boundaries and power pole positions were captured using photogrammetric methods with an absolute accuracy of+/- 0.5 metres. In rural areas, road centreline data was captured using vehicle based GPS to an absolute accuracy of +/- 5 metres. In addition to these physical features a cadastral base was also used in order to provide parcel information, both urban and rural, This will allow for the interfacing of customer databases. Portability Field inventory, by its very nature, involves taking data capture technology out of the ofllce and into the field. The methods used are many and varied, however, in most circumstances three levels of portability are required: vehicle based for normal roadside capture, vehicle based for off-road capture where the land allows for four wheel drive access, and total portability where the contour or cover allows for foot access only. In the PIMS project all three methods have been required. The equipment used obviously depends on the method, ranging from personal computers with 14 inch monitors, 12 volt hand held digital video cameras and fixed GPS for roadside capture, laptop computers, 12 volt digital cameras and portable GPS for off road capture and weather proof ruggedised laptop computers, portable digital cameras and portable GPS for capture on foot. All systems must be reliable, cost effective, freely available and physically suitable in terms of size and weight. Components have needed to be purchased off the shelf and customised to provide systems that comply with these requirements. Innovation and lateral thinking are prime attributes for this particular task. Quality Assurance The area of quality assurance is without doubt the most challenging aspect of field inventory. The principle of a single site visit and checking data quality after that site visit is obviously in conflict. The very nature of field inventory dictates that data can only be checked on site, as that is the only data source. In practice, quality assurance procedures must be matched to data quality expectations. In the PIMS project those quality expectations are very high. As a result, practices have been developed, admittedly on a trial and error basis at times, that have proved reliable and consistent, These practices fall into two broad categories: on-site and off-site. On-site quality assurance has been discussed and includes logic and plant compatibility checks via the data capture software, and positional accuracy and spatial relationships via the use of physical landbases of a defined accuracy linked to real-time differential GPS. Off-site quality assurance has seen the use of digital imagery to allow a post-capture observation of every pole structure. This image is gee-referenced to pole coordinates on-site, thus precluding any possibility of an image/pole conflict. The attributes that are capable of being checked this way are entirely dependent on the skill of the camera operator and the resolution of the image. There is an obvious desire to maximise both factors to allow for extensive re-observation. Where real-time differential GPS is impossible due to local topography or conditions, post processed differential positions are required. This precludes the availability of positional checking on-site. However, with physical landbase use and analysis of span lengths, off-site checking of post processed positions has proved reliable. In urban areas where there is a high proportion of complicated pole structures, there has been the need to compromise the single visit principle. Digital imagery in these areas has shortcomings and as a result personnel have been re-visiting some sites to confirm data. Overall, experience dictates quality assurance of field inventory data capture. Conclusion This paper has focused on using available technology in innovative ways in order to solve real problems facing utilities. Two crucial points are central to this process: ensuring the traditional attitudes between client and contractor are rejected and replaced with an environment suitable for innovation, and the fact that experience (or trial and error) is a pre-requisite to success. References: Price,M,Pickford,P, 1996,ANUFM Beyond 2000- A Practical Case Study :AM/FM International Inaugural Australasian Conference Proceedings | ||
|
|