KIMV Project
| Project Title | KIM Validation for EO archived data exploitation support | |
| Project Acronym | KIMV | |
| Contractor(s) | ACS, DLR, ETHZ |
|
Context |
| Objectives | |
| Architecture | |
| How it Works | |
| Output |
| Context | 
|
In recent years our ability to store large quantities of data has greatly surpassed our ability to access and meaningfully extract information from it. The state-of-the-art of operational systems for Remote Sensing data access, in particular for images, allows queries by geographical location, time of acquisition or type of sensor. This information is often less relevant than the content of the scene, i.e.: structures, objects or scattering properties.
Emerging needs from big applications (e.g.: change detection, global monitoring, disaster management support, etc.) and the continuous increase in archives' size and EO sensors' variety, require new methodologies and tools for information mining and management, supported by shared knowledge. The traditional interpretation of EO images and products requires to large extent the use of experts' knowledge in each application area.
Through visual inspection the experts are able to extract the information embedded in the images, and to and classify and interpret it as required, usually for a single application. The manual process performed by experts to mine information from images is currently too complex and expensive to be applied systematically on even a small subset of the acquired scenes. This limits the full exploitation of the petabytes of archived or new data. The issue might become even more challenging in future since more missions - including constellations - are being planned, with broader sensor variety, higher data rates and increasing complexity. As an example, ENVISAT alone accumulates 400 terabytes of data every year. The problematic is common also to other domains, like medicine, multimedia, and to a broad spectrum of other sensors' data.
This interpretation process takes time and it is expensive, and it cannot be used for the systematic processing and classification of large data volumes. Therefore, there is a need for automated feature recognition and classification techniques to replace the manual interpretation activity. Eventually these techniques should support the recognition of image features by automatic or semiautomatic scene analysis and classification. Results from current R&D activity might ease the access to the imagery (today mostly retrieved using spatio-temporal and a few more parameters, referred to in the following as spatio-temporal-parameter) also through their information content. The need to access information also in large volumes of image data has stimulated the research in the field of content-based image retrieval during last decade.
Many new concepts have been developed and prototyped. However the dramatic increase in volume, details, diversity and complexity, and the user demand for simultaneous access to multi-domain data urgently requires new approaches for image information mining, multi-domain information management, and knowledge management and sharing (in support to information mining and training).
Recently ESA has funded research in the field trough two TRP projects:
|
KIM: Knowledge Driven Information Mining in Remote Sensing Image Archives | |
|
|
KES: EO domain specific Knowledge Enabled Services |
The KIM / KES prototype technique for information mining differs from traditional feature extraction methods (analysing pixels and looking for a predefined pattern). It is based on extracting and storing basic characteristics of image pixels and areas, which are then selected (one or more and weighted) by users as representative of the searched feature. This method has a number of advantages:
|
|
No need to re-scan the entire image archive when searching new features | |
|
|
The selected feature can be closer to user expectations and perception (the same feature can have different meanings to different users: e.g. a forest for an environmentalist, a forest guard, a geologist, a urban planner, …) | |
|
|
The system can be implemented to learn from experts' knowledge |
| Objectives |
|
Here, KIM is meant to be the prototype developed within the KIM TRP project, but upgraded in the ingestion part within the KES TRP project.
The scope of KIMV is, starting for KIM, to implement and test a quasi-operational environment for simple access also as MASS Services to enhanced image selection functions (selection of images through a combination of standard spatio-temporal-parameter queries and image information content queries).
This requires to fulfil the following objectives:
|
|
Identify the key features and related data sets and applications or services likely to benefit at most from enhanced image selection | |
|
|
Identify and implement related KIM upgrade requirements | |
|
|
Implement the interfaces with an existing archive, via disk storage or off-lime media | |
|
|
Ensure easy operability and adequate performance of the data ingestion chain in a quasi-operational environment | |
|
|
Implement a generalised approach to envelope within the MASS environment the available functions as services for simple user access to enhanced image selection | |
|
|
Evaluate the performance of the system and its capability to fulfil user expectations in terms of image selection results (positive and negative) |
The major KIM upgrade requirements are:
|
|
Extend the supported EO sensors and image features | |
|
|
Link Image Information Mining to a specific area of user interest | |
|
|
Include surface shape as one of the Image Information Mining elements | |
|
|
Permit high throughput data ingestion within a scalable Linux Cluster | |
|
|
Combine standard catalogue search parameters (sensor, date, time, area) with Image Information Mining capabilities (feature identification) | |
|
|
Permit the extraction of geometrical objects or thematic maps | |
|
|
Enhance the image selection client (user / expert interface) | |
|
|
Envelop system functions for provision of simple services through MASS | |
|
|
Implement the MASS Gateway for recombination of internal image information content search results with internal or external spatio-temporal-parameter search results | |
|
|
Implement incremental clustering in order to permit continuous loading of new images (clustering is a technique used to compress the space required for internal storing of pixel related information) | |
|
|
Implement methods supporting fast search for improving retrieval times for large data sets | |
|
|
Implement near real-time ingestion for very fast loading of single images, by exploiting all the parallel processing capacity offered by the Linux Cluster |
| Architecture |
|
Figure 1 - Concept view of KIMV within the surrounding environment
| How it Works |
|
Figure 1 provides an overview of the KIMV concept, based on the enhanced KIM prototype. The data ingestion part shall permit sustained loading with minimum operator intervention of large image data sets from CD-ROMs or on-line storage provided by an attached archiving facility. This requires the use of automated CD-ROM reading and a Redundant Array of Inexpensive Computers (RAIC), in order to decrease ingestion time and to take advantage of scalable processing power, load balancing and fault tolerance.
At the other end KIMV shall be interconnected to MASS via the Tool-box in order to permit enhanced image selection Services, that is Services permitting to identify with simple interactions the images also through their information content (presence or absence of specific features like cloud cover, algae blooms, …). The KIM Gateway shall dynamically combine the two types of search results (those deriving from standard spatio-temporal-parameter attributes like sensor, area, time, …, and those deriving from image search via content) before providing the output to MASS.
In order to cope with the practical case of KIMV (hosting a subset of the images) attached to an archiving centre (hosting all the images and related full inventory), the KIM Gateway shall be implemented in a generic way, in order to accept results of standard spatio-temporal-parameter queries also from external sources, and to recombine them with internal results of image information content queries.
| Output |
|
The capability of KIMV to handle for example the following data sets / features will be assessed:
|
|
Landsat, MERIS, and NOAA AVHHR quick-looks over Europe for selecting cloud-free images and images showing snow cover | |
|
|
Full resolution Landsat images and ERS SAR GEC products over Europe for the identification of land features | |
|
|
Quick-look or L1b full or reduced resolution MERIS images over seas for the identification of potential algae blooms | |
|
|
ERS SAR or ENVISAT ASAR quick-look and full resolution images over seas for the identification of sea features |
The Final Presentation was held at ESRIN on April 6, 2005. At the same occasion also the report on KIM Validation was presented.
|
|
