**4. Processing the scanned data**

Dimensions and volumes of logs corresponding "exactly" to their geometrical properties cannot be, naturally, determined in operation. Every determination of the log volume based on the calculation of the volume of geometrical bodies (cylinder, truncated cone, paraboloid etc.) represents only approximation to reality but not its expression. A value achieved by this way represents a nominal "commercial" volume of wood. According to the technical feasibility of measurements under given conditions, availability of means necessary for measuring, requirements and experience of the result users and according to "historical" usage in the given region, many methods of measurements and processing the scanned data are used in Central Europe at present. Any used procedure cannot be considered to be unequivocally bad. Thus, unification of the method of wood measurement, at least in the area of Central Europe, does not appear to be realistic. Therefore, it is necessary at least to recalculate results achieved by particular procedures with each other. Unambiguous recalculation is not possible because at processing results of measurements it is necessary to use statistical methods. Under conditions of Central Europe, the volume of logs is usually given in m3 under bark (u.b. volume). It is calculated as the volume of a cylinder, the length of which is equal to the nominal length of logs and diameter corresponds to the mid diameter of logs (u.b.).

The log volume calculation as the volume of a cylinder stems from a historical method, which was proposed by a Bavarian forest inspector Franz Xaver Huber in 1825 on the basis of his theoretical analyses and long-time experience. Owing to the simple feasibility of measurements and the method of calculation as well as due to the sufficient accuracy of results, the method was soon used not only in Bavaria but in whole Germany, Austria-Hungary, and in the course of time also in a lot of other European countries.

Simplification of the log form causes that the method of Huber gives generally lower volume compared to reality providing satisfactory results only for the wide average of the large number of logs but not for particular logs. Accuracy of the measurement depends

revolution gives the number of pulses per the conveyer line unit. A log moving on a conveyer cuts across the photocell ray and during its shading the photocell sends a signal. The number of pulses sent by the generator during the photocell shading gives the log length. Accuracy reached by this method of measurement ranges from ±1 to ±2 cm. Advantages of this method consist in its simplicity, reliability and non-sensitivity to the conveyer speed and its changes during measurement. Disadvantages consist in the photocell sensitivity to defects on the log end (chamfer cut, torn up fibres). Thanks to this fact, the system has a tendency to give excess values at logs with these defects. Erroneous

Data on dimensions and shape of logs are also provided by scanners working on the principle of absorption of microwave radiation. Systems equipped with these scanners are primarily determined for scanning quality. Many defects (rot, knots, and growth anomalies) are often not visible from the stem surface being not noted at the visual checking the quality. Evaluation of scanned data at this method of scanning is however, substantially more complicated with respect to mutual relationships of more values (dimensions, density, moisture etc.). Thus, they

Dimensions and volumes of logs corresponding "exactly" to their geometrical properties cannot be, naturally, determined in operation. Every determination of the log volume based on the calculation of the volume of geometrical bodies (cylinder, truncated cone, paraboloid etc.) represents only approximation to reality but not its expression. A value achieved by this way represents a nominal "commercial" volume of wood. According to the technical feasibility of measurements under given conditions, availability of means necessary for measuring, requirements and experience of the result users and according to "historical" usage in the given region, many methods of measurements and processing the scanned data are used in Central Europe at present. Any used procedure cannot be considered to be unequivocally bad. Thus, unification of the method of wood measurement, at least in the area of Central Europe, does not appear to be realistic. Therefore, it is necessary at least to recalculate results achieved by particular procedures with each other. Unambiguous recalculation is not possible because at processing results of measurements it is necessary to use statistical methods. Under conditions of Central Europe, the volume of logs is usually given in m3 under bark (u.b. volume). It is calculated as the volume of a cylinder, the length of which is equal to the nominal length of logs and diameter corresponds to the mid

The log volume calculation as the volume of a cylinder stems from a historical method, which was proposed by a Bavarian forest inspector Franz Xaver Huber in 1825 on the basis of his theoretical analyses and long-time experience. Owing to the simple feasibility of measurements and the method of calculation as well as due to the sufficient accuracy of results, the method was soon used not only in Bavaria but in whole Germany, Austria-

Simplification of the log form causes that the method of Huber gives generally lower volume compared to reality providing satisfactory results only for the wide average of the large number of logs but not for particular logs. Accuracy of the measurement depends

Hungary, and in the course of time also in a lot of other European countries.

measurements are also caused by the log shift on a conveyer.

**4. Processing the scanned data** 

diameter of logs (u.b.).

are not used only for scanning dimensions due to their demands and costs.

mainly on the stem part, which was used for the log production. The method undervalues the volume of butt logs while top logs are overvalued. F.X. Huber was aware of this fact and, thus, for more accurate measurements, he recommended a section method. This method divides a stem to 1 or 2 m long sections and the volume of each of the sections is calculated separately using the way described above. The volume of a log is then the sum of volumes of particular sections. At the time of its origin, the method was used only for research operations due to its excessive time consumption. Due to similar causes even other procedures requiring more measurements were not used later in practice.

Procedures used at present for wood measurement and determination of its volume (from the aspect of their user in Czech Republic) are as follows:


Besides, there is a European standard EN 1309-2 Roundwood and sawn timber – Methods of measuring dimensions - Part 2: Roundwood – Requirements for measurements and rules for the volume calculation, 2006. Although it concerns a relatively new European standard, its use has not been found out in the CR. Its use (anywhere) limits not quite unambiguously determined methods of the determination of a mid diameter. A normative supplement B evokes also certain confusion. The normalized procedure of measurement is presented here as "rules for the measurement and calculation of the log volume valid if there are no state, regional or district rules".

For users in Czech countries, it is suitable to include (from practical aspects) the CSN 48 0050 Standard "Raw timber. Basic and common provisions" into the survey of rules (ČSN 48 0050, 1992). This standard was used in the Czech Republic until the publication of "Recommended rules 2002". At present, it is not legally binding being virtually not used for timber reception. The majority of users are accustomed to results of measurements carried out according to the standard. The users compare often values obtained according to other rules with its results. Moreover, results of measurements carried out according to this standard were in very good agreement with reality.

The analysis of methods of measurements includes all regulations and rules mentioned above talking into account also variants, which are determined or admitted by these regulations. It refers to following variants:


The way of carrying out particular steps is given in table 1. Variants of the implementation of particular steps (if they are determined or made possible by regulations) are given as

A step, which is consistently ignored by all regulations, is filtration of data. The aim of filtration is to create the stem image, which approaches maximally its actual/real form. It removes error data replacing them by probable data. Extreme values are considered to be "erroneous". These data originate usually by tattered parts of bark, wood, remains of branches etc. These values are replaced (at filtration) by values, which level the form of a log

In addition to defects on the wood surface, filtration is inevitable at systems with a throughway conveyer to filtrate parts of the conveyer (usually driving dogs/carriers and their guide). At manual measurements, a principle can be considered to be a certain form of filtration, which says that if there is a defect in the place of measurement (e.g. in the log centre), two measurements are carried out at both sides of the defect where the defect already does not appear. An average value from both these measurements is considered to be a value in the

Methods of filtration at electronic measurements are based on mathematical and statistical procedures, which are usually combined. Basic used procedures usually are as follows:

 *Moving averages* – serve for the general adjustment of the stem surface curve. To each of the places of measurement (x) a value is assigned calculated as an average from values taken in an interval (x - n, ... x, ... x + n). In its centre, there is the place of a given measurement. The number of members serving for the calculation of a moving average (smoothing with) is usually odd. The procedure can be also several times repeated (the depth of smoothing), however, there is a danger of too large idealization of the piece (log) form. By means of this method, it is not possible to calculate values for edge places of measurement. At the determination of the log mid diameter for the calculation of its volume, the shortcoming does not appear but when we need to determine top diameter

 *Moving medians* – serve also for the general adjustment. The procedure is anagogical the previous method, only the average value is replaced by median. Advantages consist in

 *Top extremes cutting* – serve to eliminate extreme values with positive deviation caused usually by protruding bark, torn up fibres or parts of branches. More types of methods of the elimination of positive extreme values are used. Nevertheless, they do not suffice

 *Mutual comparison of values of diameters measured in perpendicular directions* (e.g. horizontally - x and vertically - y) in one place with respect to the log length. Differences in values X and Y are compared with a value higher than common flattening. Effects of flattening can be eliminated comparing several successive values of

alone for the filtration, always follows adjustment using some of other methods. *Linear regression –* serves for the total adjustment of the stem surface curve to a straight line by the approximation of given values using the polynomial of the first order (straight line) by the method of least squares. If this procedure is applied directly to measured values, considerably distant values can markedly affect (deviate) the whole line. Its use for the set of values, which were already adjusted by mean of another

in such a way to correspond (ie with higher probability) its real form.

it is necessary to extrapolate it in the course of moving averages.

fact that the resulting value is not affected by potential extremes.

separate procedures.

original place of measurement.

method, was more suitable.

 The CSN 48 0050 Standard "Raw timber. Basic and common provisions" determines different procedures for manual and electronic measurements.

A detailed analysis in the previous determination of procedures and their variants (which does not include all European regulations) has shown that it is possible to analyse them to common elementary steps. Not all steps prescribe all procedures and the method of implementing many steps is different. However, by the exact definition of particular steps, all procedures can be unambiguously and fully characterized (see Fig. 7).

Fig. 7. Elementary steps at scanning and evaluation of dimensions and volume of logs and their sequence.

The CSN 48 0050 Standard "Raw timber. Basic and common provisions" determines

A detailed analysis in the previous determination of procedures and their variants (which does not include all European regulations) has shown that it is possible to analyse them to common elementary steps. Not all steps prescribe all procedures and the method of implementing many steps is different. However, by the exact definition of particular steps,

> *Length scanning [cm]*

*Evaluation of nominal length [cm]* 

Fig. 7. Elementary steps at scanning and evaluation of dimensions and volume of logs and

their sequence.

different procedures for manual and electronic measurements.

all procedures can be unambiguously and fully characterized (see Fig. 7).

*Diameter scanning [mm]* 

*Selection of places of measurement* 

*Data filtration* 

*Conversion of diameter values from mm to cm*

*Selection of diameter values* 

*Determination of the diameter value in the place of measurement*

*The conversion of diameter values in the place of measurement from mm to cm*

*Selection of places of measurement used for the calculation of mid diameter* 

*Determination of mid diameter values*

*Conversion of mid diameter values*

*Bark deduction (if measurements are carried out over bark)*

*Log volume calculation* 

The way of carrying out particular steps is given in table 1. Variants of the implementation of particular steps (if they are determined or made possible by regulations) are given as separate procedures.

A step, which is consistently ignored by all regulations, is filtration of data. The aim of filtration is to create the stem image, which approaches maximally its actual/real form. It removes error data replacing them by probable data. Extreme values are considered to be "erroneous". These data originate usually by tattered parts of bark, wood, remains of branches etc. These values are replaced (at filtration) by values, which level the form of a log in such a way to correspond (ie with higher probability) its real form.

In addition to defects on the wood surface, filtration is inevitable at systems with a throughway conveyer to filtrate parts of the conveyer (usually driving dogs/carriers and their guide).

At manual measurements, a principle can be considered to be a certain form of filtration, which says that if there is a defect in the place of measurement (e.g. in the log centre), two measurements are carried out at both sides of the defect where the defect already does not appear. An average value from both these measurements is considered to be a value in the original place of measurement.

Methods of filtration at electronic measurements are based on mathematical and statistical procedures, which are usually combined. Basic used procedures usually are as follows:


The method of filtration shows substantial effects on values of quantities evaluated in next steps. Different methods of filtration are selected both for various types of sensing devices and assortments of processed raw material. Similarly, it is possible to achieve the required conformity of results of standard measurements with the given reference measurement by the suitable method of filtration. Therefore, manufacturers of sensing (scanning) or control systems use frequently filtration at the final setting the equipment parameters. Thus, it is not

methods of the electronic scanning of wood provide very similar data on its geometrical

 methods of the evaluation of electronically scanned data markedly follow the manual method of measurement trying to adapt to its possibilities (number of used

 methods of determination of the stem (log) volume keep the original principle of the log volume calculation as one cylinder although the density of carried out measurements provides the sufficient number of data for the calculation of the log volume according to

Thus, the absence of using possibilities of electronic scanning and evaluation of dimensions, form and volume of wood results in deviations between evaluated (nominal, commercial)

Differences in the determination of dimensions (above all mid diameter) of logs between particular rules result in differences in the determination of wood volume according to

conversions generally (only exceptional using millimetre values for next calculations),

 the number of conversions (the conversion of particular scanned values and values calculated as an average from values conversed to whole cm already previously), determination of the place of measuring the mid diameter of logs (centre of the nominal

 evaluation of the log mid diameter (average value or a smaller value from diameter values determined in both places of measurement within the measuring area in the

 selection of diameter values (keeping values taken horizontally and vertically /2D equivalent of scanning/ or searching for a maximum) – only at 3D scanning, filtration of scanned data (its effect at present valid regulations cannot be determined

Results of measurements and determination of the volume of logs obtained according to rules given in table 1 are compared with the value of a volume, which approaches most the geometrical volume of logs. Calculation of the "geometrical" volume of logs is based exclusively on values scanned on a regular basis according to present regulations. Thus,

particular rules/regulations. Steps, which cause these differences, are as follows: conversion of millimetre values of the log diameter to whole centimetres,

the method of conversion (mathematical rounding or removing units in mm),

because the regulations do not define the way of its implementation).

or actual geometrical length including allowances),

possible to define concrete used methods of filtration.

properties,

sections.

measurements, accuracy),

volume and real/actual volume.

centre of the log length),

**5. Differences in results** 

The analysis of particular steps results in the following conclusions:

the X and Y difference. The flattening becomes evident in the whole length of the log or at least on its longer section. The procedure serves to identify "unreliable values". For their replacement, it is necessary to use another method.


To record the majority of described erroneous data but, at the same time, for the realistic implementation of filtration there are usually more suitable combined methods. It is possible to combine both various types of basic methods and their succession and in "moving methods" also the number of values taken into account. At the same time, different methods are used for the filtration of the log central part to obtain mid diameter for the volume calculation (or for cross-cutting) and for the filtration of the log end parts to determine end diameter mainly for the purpose of grading. There, the use of "moving" procedures is very limited.

Fig. 8. An example of the effect of various procedures of filtration of scanned data on the resulting image of a log in a control computer (the log shows ragged bark in its central part).

 *Mutual comparison of values of diameters measured successively in one direction.* A difference between Xa and Xa+1 is compared with the maximum admissible size of a difference derived from the possible stem taper. Similarly as the previous method it serves to identify "unreliable values". The method disadvantage is that it does not find an error which consists in the deviation of more successive values. It is not also utilizable for

 *Comparison of the growth coefficient of successive values,* i.e. the relation of Xa and Xa+1 with a value derived from a possible taper. Members of the proportion may not be successive measured values. Also this method serves to identify defects and also here it is necessary to determine another method or to change the comparison of data at the end

To record the majority of described erroneous data but, at the same time, for the realistic implementation of filtration there are usually more suitable combined methods. It is possible to combine both various types of basic methods and their succession and in "moving methods" also the number of values taken into account. At the same time, different methods are used for the filtration of the log central part to obtain mid diameter for the volume calculation (or for cross-cutting) and for the filtration of the log end parts to determine end diameter mainly for the purpose of grading. There, the use of "moving"

their replacement, it is necessary to use another method.

searching defects at the log end.

of a series.

procedures is very limited.

0

0

200

390

590

790

980

1180

1370

1570

1760

1960

X median Xm median X and Y

Fig. 8. An example of the effect of various procedures of filtration of scanned data on the resulting image of a log in a control computer (the log shows ragged bark in its central part).

Y median Ym diameter Xm and Ym

2160

**Log length [mm]**

2350

2550

2740

2940

3140

3330

3530

3720

3920

4020

50

100

150 200

250 300

**Log diameter [mm]**

350 400

450

the X and Y difference. The flattening becomes evident in the whole length of the log or at least on its longer section. The procedure serves to identify "unreliable values". For The method of filtration shows substantial effects on values of quantities evaluated in next steps. Different methods of filtration are selected both for various types of sensing devices and assortments of processed raw material. Similarly, it is possible to achieve the required conformity of results of standard measurements with the given reference measurement by the suitable method of filtration. Therefore, manufacturers of sensing (scanning) or control systems use frequently filtration at the final setting the equipment parameters. Thus, it is not possible to define concrete used methods of filtration.

The analysis of particular steps results in the following conclusions:


Thus, the absence of using possibilities of electronic scanning and evaluation of dimensions, form and volume of wood results in deviations between evaluated (nominal, commercial) volume and real/actual volume.

Differences in the determination of dimensions (above all mid diameter) of logs between particular rules result in differences in the determination of wood volume according to particular rules/regulations. Steps, which cause these differences, are as follows:

