Filtern
Erscheinungsjahr
Dokumenttyp
- Zeitschriftenartikel (21) (entfernen)
Schlagworte
Organisationseinheit der BAM
The preservation of human achievements and progress is undoubtedly a pillar of our society. This includes detecting threats in time to prevent disasters that may occur from forces of nature, defective technology, human failure or malicious intent. In this respect, there is arguably no difference between safety and security, though different institutions are involved in the regulation and control of them. However, the unifying rationale ought to result in the building of bridges between them. This is the Working Group's mission.
An algorithm has been developed to remove reliably dotted interferences impairing the perceptibility of objects within a radiographic image. This particularly is a major challenge encountered with neutron radiographs collected at the NECTAR facility, Forschungs-Neutronenquelle Heinz Maier-Leibnitz (FRM II): the resulting images are dominated by features resembling a snow flurry. These artefacts are caused by scattered neutrons, gamma radiation, cosmic radiation, etc. all hitting the detector CCD directly in spite of a sophisticated shielding. This makes such images rather useless for further direct evaluations.
One approach to resolve this problem of these random effects would be to collect a vast number of single images, to combine them appropriately and to process them with common image filtering procedures. However, it has been shown that, e.g. median filtering, depending on the kernel size in the plane and/or the number of single shots to be combined, is either insufficient or tends to blur sharp lined structures. This inevitably makes a visually controlled processing image by image unavoidable. Particularly in tomographic studies, it would be by far too tedious to treat each single projection by this way. Alternatively, it would be not only more comfortable but also in many cases the only reasonable approach to filter a stack of images in a batch procedure to get rid of the disturbing interferences.
The algorithm presented here meets all these requirements. It reliably frees the images from the snowy pattern described above without the loss of fine structures and without a general blurring of the image. It consists of an iterative, within a batch procedure parameter free filtering algorithm aiming to eliminate the often complex interfering artefacts while leaving the original information untouched as far as possible.
The effective high neutron scattering absorption coefficient of hydrogen (48.5 cm²/g) due to the scattering allows neutrons to reveal hydrocarbon structures with more contrast than X-rays, but at the same time limits the sample size and thickness that can be investigated. Many planar shaped objects, particularly wood samples, are sufficiently thin to allow thermal neutrons to transmit through the sample in a direction perpendicular to the planar face but not in a parallel direction, due to increased thickness. Often, this is an obstacle that prevents some tomographic reconstruction algorithms from obtaining desired results because of inadequate information or presence of distracting artifacts due to missing projections. This can be true for samples such as the distribution of glue in glulam (boards of wooden layers glued together), or the course of partially visible annual rings in trees where the features of interest are parallel to the planar surface of the sample. However, it should be possible to study these features by rotating the specimen within a limited angular range. In principle, this approach has been shown previously in a study with fast neutrons [2]. A study of this kind was performed at the Antares facility of FRM II in Garching with a 2.6×107/cm² s thermal neutron beam. The limit of penetration was determined for a wooden step wedge carved from a 2 cm×4 cm block of wood in comparison to other materials such as heavy metals and Lucite as specimens rich in hydrogen. The depth of the steps was 1 cm, the height 0.5 cm. The annual ring structures were clearly detectable up to 2 cm thickness. Wooden specimens, i.e. shivers, from a sunken old ship have been subjected to tomography. Not visible from the outside, clear radial structures have been found that are typical for certain kinds of wood. This insight was impaired in a case where the specimen had been soaked with ethylene glycol. In another large sample study, a planar board made of glulam has been studied to show the glued layers. This study shows not only the limits of penetration in wood but also demonstrates access to structures perpendicular to the surface in larger planar objects by tomography with fast neutrons, even with incomplete sets of projection data that covers an angular range of only 90° or even 60°.
Stochastic artefacts are frequently encountered in digital radiography and tomography with neutrons. Most obviously, they are caused by ubiquitous scattered radiation hitting the CCD-sensor. They appear as scattered dots and, at higher frequency of occurrence, they may obscure the image. Some of these dotted interferences vary with time, however, a large portion of them remains persistent so the problem cannot be resolved by collecting stacks of images and to merge them to a median image. The situation becomes even worse in computed tomography (CT) where each artefact causes a circular pattern in the reconstructed plane. Therefore, these stochastic artefacts have to be removed completely and automatically while leaving the original image content untouched. A simplified image acquisition and artefact removal tool was developed at BAM and is available to interested users. Furthermore, an algorithm complying with all the requirements mentioned above was developed that reliably removes artefacts that could even exceed the size of a single pixel without affecting other parts of the image. It consists of an iterative two-step algorithm adjusting pixel values within a 3 × 3 matrix inside of a 5 × 5 kernel and the centre pixel only within a 3 × 3 kernel, resp. It has been applied to thousands of images obtained from the NECTAR facility at the FRM II in Garching, Germany, without any need of a visual control. In essence, the procedure consists of identifying and tackling asymmetric intensity distributions locally with recording each treatment of a pixel. Searching for the local asymmetry with subsequent correction rather than replacing individually identified pixels constitutes the basic idea of the algorithm. The efficiency of the proposed algorithm is demonstrated with a severely spoiled example of neutron radiography and tomography as compared with median filtering, the most convenient alternative approach by visual check, histogram and power spectra analysis.
Wood is the material that has accompanied the whole development of mankind in various applications, for manufacturing tools and weapons, for buildings and constructions and also as fuel. It has various appearances and is subjected to decomposing changes, so there are sufficient arguments for non-destructive testing of wooden objects in the same way as is common practice with other technologically used materials. However, even today wood is rarely tested. Moreover, artefacts of cultural heritage containing wood are rare and delicate, so dismantling these for studying purposes is undesirable. Radiological inspection technologies are the methods of choice to provide a non-destructive insight. Interrogations of a specimen with various kinds of radiation provide different answers about their internal structures. High-quality detailed images are achievable with X-rays or gamma radiation in cases where heavy metals or other dense materials are not involved. Neutrons behave complementarily; they are avidly absorbed by light elements such as hydrogen on the one hand and yet are capable of easily penetrating heavy metals on the other. This provides an alternative for X-ray radiography and tomography when material characteristics are of primary interest rather than structural details, or when shielding with plates or sleeves of heavy metal severely impedes inspections with X-ray or gamma radiation technologies. However, due to the moderating effect of wooden samples it is essential to use fast neutrons for radiography and tomography of voluminous objects. Some typical examples described here will show the difference between neutron and X-ray photon-based radiographic technologies.