Digitisation

Almost all modem laboratory based data are now obtained via computers, and are acquired in a digitised rather titan analogue form. It is always important to understand how digital resolution influences the ability to resolve peaks. 

Many techniques for recording information result in only a small number of data-points per peak. A typical NMR peak may be only a few hertz at half-width, especially using well resolved instrumentation. Yet a spectrum recorded at 500 MHz, where 8K (=8192) datapoints are used to represent 10 ppm (or 5000 Hz) involves each datapoint 

Consider a Gaussian peak, with a true width at half-height of 30 units, and a height of 1 unit. The theoretical area can be calculated using the equations in Section 3.2.1.1  

Typical units for area might be AU.s if the sampling time is in seconds and the intensity in absorption units. 

Influence on the appearance of a peak as digital resolution is reduced 

---

After often a lengthy period (several months) of acquisition and processing, the data may be loaded onto a seismic workstation for interpretation. These workstations are UNIX based, dual screen systems (sections on one side, maps on the other, typically) where all the trace data is stored on fast access disk, and where the picked horizons and faults can be digitised from the screen Into a database. Of vital Importance is access to all existing well data in the area for establishing the well - seismic tie. 2D data will be interpreted line by intersecting line, and 3D as a volume. 

Maps can be created by hand or by computer mapping packages. The latter has become standard. Nevertheless, care should be taken that the mapping process reflects the geological model. Highly complex areas may require considerable manual input to the maps which can subsequently be digitised. 

Undercuts and cracks are represented in the digitised radiograph as local greyvalue minima (see Fig. 2). This motivates the application of edge-detecting operators. 

Projection radiography is widely used for pipe inspection and corrosion monitoring. Film digitisation allows a direct access to the local density variations by computer software. Following to a calibration step an interactive estimation of local wall thickness change based on the obtained density variation is possible. The theoretical model is discussed, the limitations of the application range are shown and examples of the practical use are given. The accuracy of this method is compared to results from wall thickness measurements with ultrasonic devices. 

For the practical evaluation of the algorithm described previously it is integrated into the NDT Sean Manager system (DBA Systems Inc, Melbourne, FL, U.S.A). This system allows film digitisation, display, evaluation and archiveraent of images /3,4/ and was developed for the needs of computer based industrial NDT film inspection. A snapshot of the user interfaee for wall thickness evaluation is shown in fig. 3. 

Fig. 5 Erosion pit inside a reducing pipe fitting, projection technique at 160 kV, profile plot with optical densities of the digitised film. The varying background caused by the geometrical set-up prevents a wall thickness calibration as in fig. 4...

Sample Frequency. Defines the sampling frequency, in MHz, to be used in the digitisation of the A-scan signal. 

The corrosion process is observed as a series of events which all contribute to the overall corrosion rate. Measurement of rest potential fluctuations between two identical electrodes of potential fluctuations with respect to a fixed reference can be carried out. The electrochemical noise output spectrum is analysed using digitised data. The interpretation requires electrochemical expertise, and the method is therefore usually provided as a specialised service. 

Figure 7.4. Digitised cross sections of femora from 67-day-old rats (a) normal control and (b) four-week suspended animals. Cross-sectional area of (a) is 5.3 mm and (b) is 3.8 mm, a 29 per cent reduction due to the unloading during growth.

Decades of combined spectral and chemistry expertise have led to vast collections of searchable user databases containing over 300 000 UV, IR, Raman and NMR spectra, covering pure compounds, a broad range of commercial products and special libraries for applications in polymer chemistry (cf. Section 1.4.3). Spectral libraries are now on the hard disks of computers. Interpretation of spectra is frequently made only by computer-aided search for the nearest match in a digitised library. The spectroscopic literature has been used to establish computer-driven assignment programs (artificial intelligence). 

Unfortunately these devices are quite sophisticated and expensive and they have not been used for behaviour research. In order to assess the quality of movements and human expression we developed a system which can be used even in labs with a low budget. We dispensed with real time operation capabilities and turned to the analysis of digitised video. These procedure is called Automatic Movie Analysis (AMA). The advantages are clear it is possible to repeat any type of analysis and control for artefacts. 

Digitising of a movie in a range between 12.5 to 25 pictures a second (320 x 240 pixels frame size) in greyscales (Pixelvalue=[0..256] greys, where 0 is white and 256 is black). 

---