Method for detecting wood defects and wood defect characteristics during the processing of tree trunks and device for processing tree trunks

Abstract

To enable reliable quality inspection and high timber yield when processing a tree trunk, and to minimize metrological labor at the same time SOLUTION: A method for detecting timber defects and timber defect characteristics when processing a tree trunk, comprising the following method steps: A) performing a profile measurement of the tree trunk (1); Determining a trunk geometry of the tree trunk, B) carrying out a separation processing of the tree trunk (1) with formation of a processing surface (8) along a trunk axis of the tree trunk (1), C) carrying out an imaging inspection of the separated tree trunk (1) at least at the processing surface (8) and determining at least one inspection image (14), D) Spatially resolved determination of timber defects (10) and / or timber defect properties in the separately processed tree trunk as a function of the inspection image (14) according to method step C) and the trunk geometry according to method step A).SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a method for detecting timber defects and / or timber defect characteristics when processing tree trunks.Furthermore, the present invention relates to an apparatus for processing tree trunks. [Background technology]

[0002] In industrial timber processing, tree trunks are processed, among other things, into planks. The goal is to obtain the highest possible timber yield. To achieve this, it is generally necessary to determine how the planks to be obtained should be arranged relative to the trunk's shape and what their dimensions should be in order to obtain the desired high timber yield. To do this, the trunk is first scanned using a profile sensor. The three-dimensional trunk shape and dimensions are determined based on the measurement data obtained, and the arrangement and dimensions of the planks to be obtained are determined based on these. The so-called main product planks are generally located in a cross-sectional area that is on the inside of the trunk, while the so-called by-product planks are located in a cross-sectional area that is on the outside of the main product planks.

[0003] In order to obtain the by-product boards, the profiling of the by-product boards is often carried out directly on the trunk, which is generally a so-called log that has a substantially round cross section due to the natural growth of the trunk. To profile the by-product boards, at least one bark region of the trunk is first removed to form at least one working surface. In this separated and processed state, the trunk may have not only one but multiple working surfaces distributed around its periphery, for example, as a so-called model with working surfaces on two sides, or as a timber with working surfaces distributed around its four peripheries.

[0004] After the trunk is separated, it is typically scanned again with a profile sensor to determine its geometry. The collected geometry information about the trunk being separated can be used to spatially resolve wood defects, such as cracks or branch passages extending through the trunk. Post-separation profile measurements can also be used to verify that the processed surface on the trunk is formed in the desired position and orientation. This allows for the size and placement of the main and by-product boards to be obtained for the separated trunk after separation, or allows for modifications to the original cutting method so that the trunk can be processed accordingly. Summary of the Invention [Problem to be solved by the invention]

[0005] The disadvantage of measuring the profile of a separated and processed tree trunk is that it requires a lot of metrological effort. Therefore, an object of the present invention is to enable reliable quality inspection and high timber yield when processing tree trunks, while at the same time minimizing the metrological effort. [Means for solving the problem]

[0006] The problem is solved by a method according to claim 1. Advantageous developments are the subject of the dependent claims. The invention is also solved by a device according to claim 12.

[0007] The method according to the invention is used for detecting timber defects and timber defect characteristics during processing of tree trunks and comprises the following method steps: A) performing a profile measurement of the trunk and determining the shape and dimensions of the trunk; B) forming at least one processing surface along the trunk axis of the trunk and then performing a separating process on the trunk; C) performing an imaging inspection on at least the processed surface of the separated trunk to detect at least one inspection image; D) A step of spatially resolving determination of timber defects and / or timber defect characteristics in the separated and processed tree trunks depending on the inspection image according to method step C) and the geometry and dimensions of the trunk according to method step A).

[0008] The present invention is based on the realization that it is not necessary to collect information about the trunk geometry both before and after the trunk is separated. Rather, the inventive concept is based on the fact that information about the quality, such as the location and spatial extent of timber defects, can be obtained on the basis of, on the one hand, the trunk's own pre-separation geometry and, on the other hand, relatively easily obtained image data of the separated trunk. This makes it possible to determine, contrary to standard practice, in what arrangement and dimension the planks to be obtained should be arranged relative to the separated trunk in order to obtain the highest possible timber yield and timber quality without carrying out profile measurements of the separated trunk.

[0009] In a possible embodiment, the trunk is moved in a feed direction substantially along the trunk axis to perform at least one of the method steps A), B), C), and D). It is also possible that the trunk axis may be linear or may have at least one or more curvatures due to the natural growth of the trunk. Thus, the working surface of the trunk may have a substantially linear or curved extension along the trunk axis.

[0010] During method step A), the trunk is preferably a log, having a substantially round cross-sectional extension due to natural growth on the periphery and along the trunk axis of the trunk. Method step A) is preferably performed by one or more profile sensors, each of which may be configured, for example, as a laser or light section sensor or similar measuring means, and which are used to determine the geometrical profile of the trunk. It is within the scope of the present invention that the profile sensor and the trunk are in a state of relative movement relative to one another, in which case the profile sensor is particularly stationary and the trunk is transported in the feed direction during the feed movement. In this case, the geometrical dimensions of the trunk can be determined as a function of the collected profile information and the feed direction of the trunk.

[0011] Within the scope of the present invention, it is possible for the trunk geometry to be represented by a point cloud or similar primary data collected in method step A). ​​Within the scope of the present invention, the point cloud may represent the surface of the trunk and thereby the spatial extent of the trunk or a portion of the trunk. It is also within the scope of the present invention to generate a trunk model representing the trunk geometry based on the primary data. The trunk model may have a higher data density compared to the primary data and in particular may at least partially comprise a closed surface that substantially corresponds to the surface of the trunk in method step A). ​​In this case, the trunk model may be of the type of envelope model, which represents only the outer surface of the trunk. Similarly, the trunk model may be of the type of solid model with a virtual trunk interior.

[0012] The present invention is not limited by how method step B) is performed. It is conceivable that the separating of the trunk and the formation of the working surface are performed by one or more sawing means to separate the bark region from the trunk as a continuous section. It is also conceivable that the working surface is formed by machining the bark region into wood pieces, for example by means of a cutter head.

[0013] The imaging inspection of the separated and processed trunk according to method step C) can be carried out by one or more imaging sensors, in particular by a camera or camera system, which can represent at least the processed surface of the trunk in the captured image by two-dimensionally distributed grayscale and / or RGB color values.

[0014] In particular, the at least one image sensor can be arranged so that the separated trunk passes through a detection field of the image sensor at the processing surface of the trunk during the feed movement. In this case, the image sensor can have a measurement axis, which indicates the spatial extent of the detection field of the image sensor, along which the at least one image sensor has a depth of field, in which a sufficiently sharp inspection image of the processing surface can be formed by the image sensor.

[0015] Advantageously, the sensor is arranged during method step C) with its measurement axis substantially perpendicular to the processing surface. This can be easily achieved since the feed movement of the processed trunks is generally achieved by a plurality of feed rollers, the circumferential surfaces of which are in substantially flush contact with the processing surface. This allows the imaging sensor to be arranged with its measurement axis perpendicular to the processing surface of the processed trunks, for example by arranging it so that its measurement axis is perpendicular to the roll axis. The sensor arrangement can also be based on components of another device used to guide the processing surface of the processed trunks.

[0016] It is also conceivable to arrange the imaging sensor in its measurement axis at an angle to the work surface, in particular at an angle between 0 and 90 degrees. Due to the spatial extent of the depth of field of the imaging sensor along the measurement axis, it is possible to detect the work surface sufficiently clearly even in the case of such an angled arrangement. In particular, it is possible to compensate for trapezoidal distortions of the detected work surface by means of correction steps, such as keystone correction.

[0017] As mentioned above, in method step D) timber defects and / or timber defect characteristics are determined in a spatially resolved manner. The present invention is not limited to the type of timber defects or timber defect characteristics that are determined. However, it is important that the defects or defect characteristics can be determined in a spatially resolved manner in relation to the geometric dimensions of the processed tree trunk. In other words, the spatial location and / or extent of timber defects and / or timber defect characteristics at the processed surface are determined in the processed tree trunk. In particular, method step D) is performed independently of geometrical information about the processed surface of the processed tree trunk.

[0018] Timber defects may comprise branch channels and / or knotholes and / or splits and / or warping and / or discolouration and / or decay and / or sapwood and / or insect damage. Timber defect characteristics may represent timber defects, in particular the specification of any one of the aforementioned timber defects.

[0019] The determination of timber defects may comprise a data processing step by which the inspection images collected during the imaging inspection are evaluated. The evaluation preferably comprises image processing, in particular filtering and / or edge detection and / or segmentation and / or morphological processing and / or feature extraction and / or pattern recognition and / or pattern classification and / or texture analysis and / or histogram analysis, in order to determine the timber defects and / or timber defect characteristics. In particular, the location and / or extent of the determined timber defects can be represented on the inspection image by an error radar chart.

[0020] In an advantageous development, in method step A), a virtual trunk model is generated which at least partially represents the geometry of the trunk and has at least one profile cross-section, in which a virtual working plane is generated in the virtual trunk model, and the spatially resolved determination of timber defects is carried out by projecting an inspection image or information derived from the inspection image in the virtual working plane onto the virtual trunk model.

[0021] The aforementioned developments are based on the recognition that the trunk is processed in method step B) so that the processed surface formed can be virtually reproduced on the trunk model in method step A). ​​In other words, a virtual twin of the trunk is generated based on the trunk model, which virtual twin is based on the profile measurement in method step A) on the one hand and on the adjustment of the trunk model, particularly according to the separate processing in method step B), on the other hand. Based on the virtual twin, a quality inspection of the trunk is carried out by inserting an inspection image of the trunk or a part of it into the cut area of ​​the trunk. Visually, the virtual twin of the trunk can in principle correspond to a cylinder or hollow cylinder, which cylinder or hollow cylinder is provided with the inspection image or a part of it on a side that is flattened on one side.

[0022] In particular, instead of the inspection image, an error radar chart representing the location of timber defects in the inspection image can be projected onto the trunk model. Such an error radar chart can result from a data processing step performed prior to the inspection image. In other words, it is not necessary to visually couple the trunk model with the inspection image, or portions thereof, in order to be able to spatially resolve timber defects in relation to the geometry of the separated trunk. Rather, detection of timber defects and / or timber defect characteristics can be performed separately from determining the location and / or extent of the corresponding timber defects in the separated trunk.

[0023] Therefore, in an advantageous development, in method step D), timber defects and / or timber defect characteristics are detected based on the inspection image, and the position and / or extent of the timber defects and / or timber defect characteristics relative to the separated tree trunk are determined depending on common features between the inspection image and the trunk model.

[0024] The advantage of the aforementioned development is that in the first partial step D1) it is possible to detect timber defects and / or timber defect characteristics in a simple manner based solely on the inspection image. Only in the subsequent second partial step D2) can common characteristics be determined which are contained not only in the inspection image but also in the trunk model. This makes it possible to project the inspection image or the part of the timber containing the defect into a virtual processing plane of the trunk model with high accuracy, thereby enabling a spatially resolved determination of how the timber defects and / or timber defect characteristics are arranged in the separated and processed trunk.

[0025] In an advantageous development, at least one virtual edge is formed in the virtual working plane by the profile cross-section of the virtual trunk model, and timber defects and / or timber defect characteristics are determined spatially resolved according to the edge.

[0026] Within the scope of the present invention, the term "rounded edge" can be understood in principle to mean a geometric feature in a processed trunk that separates the processed surface from the unprocessed rounded edge region, which can be considered to be the outer peripheral region of the trunk that has a rounded geometry due to the natural growth of the trunk.

[0027] Research has shown that at least one rounded edge can be used as a reference feature for spatially resolving timber defects and / or timber defect characteristics with high accuracy by adjusting the projection of the inspection image or a part thereof onto the virtual trunk model. In particular, it is also possible to form two virtual rounded edges that define a projection area by including a virtual working surface in a virtual working plane by a profile cross-section, and to post the inspection image or information derived from the inspection image, such as an error radar chart, within the projection area.

[0028] In an advantageous development, the imaging inspection in method step C) is carried out in such a way that at least one round edge adjacent to the processing surface is detected and timber defects and / or timber defect characteristics are determined spatially resolved in method step D) depending on a comparison between the virtual round edge of the virtual trunk model and the round edge adjacent to the processing surface of the separated processed trunk.

[0029] The applicant's research has shown that a rounded edge can extend substantially along the trunk axis and thereby have a profile extension in the plane formed by the processing surface that particularly individualizes the separated processed trunk. A virtual rounded edge can likewise have an extension formed by the profile cross-section of a virtual trunk model in the virtual processing plane. The applicant's research has also shown that the extension of an actually formed rounded edge and a virtual rounded edge are generally sufficiently similar, at least in the virtual processing plane, that it is possible to adjust the inspection image on the virtual trunk model based on the extension of the rounded edge.

[0030] It is within the scope of advantageous developments that a comparison between the virtual edge of the virtual trunk model and the edge of the edge adjacent to the processing surface of the separated processed trunk is performed by a computing unit and an evaluation routine implemented in the computing unit. In particular, it is conceivable to detect each edge of the virtual trunk model by means of data processing, in particular by an analysis algorithm and / or machine learning or AI, and to determine at what position and / or size and / or orientation the inspection image or information derived from the inspection image, in particular the error radar chart, should be positioned in the virtual processing plane. A spatially resolved identification of timber defects and / or timber defect characteristics can then be performed accordingly.

[0031] In an advantageous development, the position and / or orientation of the virtual working plane on the virtual trunk model for the profile cross section is determined on the basis of the nominal rotational position of the trunk in method step B).

[0032] The aforementioned development is based on the recognition that, for processing the trunk, the trunk is generally rotated about its trunk axis before moving the trunk toward the respective separating means used to form the processing surface in method step B). The required rotational position depends on the trunk's geometry. In this regard, it is also possible to determine, based on the trunk's geometry in method step A) and the trunk's nominal rotational position in method step B), at what position and orientation should a virtual processing plane be located in the virtual trunk model in order to form a profile cross-section at the virtual processing plane. This advantageous development is also based on the recognition that the rotational position of the trunk for the separating process in method step B) can generally be adjusted sufficiently accurately so that the formed processing surface is in the desired position and orientation on the trunk. In other words, it is possible to determine the position and / or orientation of the virtual processing plane based on the recognition that there is no deviation between the nominal rotational position and the actual rotational position of the trunk during the separating process in method step B).

[0033] In one possible embodiment, the trunk is rotated about its axis between method steps A) and B) to move the trunk with the bark area to be removed into the required position for the process. The nominal rotational position of the trunk required for this purpose can be determined according to the trunk geometry in method step A), which can be converted into a corresponding control program, and the rotation and feed movements of the trunk can be performed according to the control program. For example, the control information can be used to position a virtual processing surface by the computing unit in such a way that the processing plane is substantially identical to the processing surface of the processed trunk relative to the virtual trunk geometry and thus represents the processing surface sufficiently accurately.

[0034] Due to the natural shape of the trunk and other influences in the industrial processing of the trunk, threading errors may occur in the trunk during the separating process in method step B), which may result in the processed surface being formed in a position and / or direction that is different from its desired position. In such cases, when a profile cross-section is formed on the trunk model assuming a nominal rotational position of the trunk, differences will occur between the processed surface and the rounded edge relative to its virtual counterpart on the trunk model.

[0035] As a result of the applicant's research, it has been found that after determining the extension of the machining surface or, for example, a rounded edge adjacent to the machining surface based on the inspection image, it is possible by image inspection to determine in what position and / or orientation the virtual machining plane should be in order to obtain an identical or at least similar virtual machining surface and / or machining rounded edge by the profile cross-section of the trunk model.

[0036] In an advantageous development for this purpose, the position and / or orientation of a virtual working plane in the virtual trunk model is determined depending on the working surface and / or at least one rounded edge of the inspection image. In particular, the virtual working plane is determined in the virtual trunk model by determining a virtual working surface and / or a virtual rounded edge of the trunk model that is identical or at least similar to the working surface and / or the rounded edge of the inspection image, respectively. The position and / or orientation of the virtual working plane is selected so that the virtual working plane extends through and in particular includes the determined virtual working surface and / or virtual rounded edge.

[0037] For example, a virtual machining plane can be determined by repeatedly cutting the stem model at a number of different positions and angles. The resulting cut surfaces can be compared with the machining surface and / or the rounded edge of the inspection image. In particular, a similarity criterion can be defined, and the position of the virtual machining plane can be determined according to the criterion. It is also conceivable to determine the virtual machining plane by inputting the extension of the machining surface and / or the rounded edge and the stem model into a calculation model and obtaining the position and / or orientation of the virtual machining plane, in particular in a single calculation step.

[0038] In an advantageous further development, the cutting strategy is determined or optimized as a function of the spatially resolved timber defects and / or timber defect characteristics determined in method step E).

[0039] The cutting scheme may comprise the arrangement of the planks of the main product and / or by-product and their respective dimensions relative to the geometric dimensions of the trunk. It is also possible to determine a new cutting scheme in the context of method step E) or to optimize an existing cutting scheme, particularly in the context of reoptimization. In particular, a first cutting scheme may already be determined based on the profile measurement in method step A), and method step B) is carried out based on the cutting scheme. After the separate processing and image inspection in method steps C) and D), a spatially resolved determination in method step D) can be used to determine a second cutting scheme, in which the defects in the timber and / or the defect characteristics of the timber are taken into account. In this case, areas of the trunk affected by defects may be excluded, or the dimensions and / or positions of the planks of the timber to be formed may be changed compared to the first cutting scheme. In particular, the division of the two by-product planks with respect to their relative positions and their dimensions may be changed compared to the first cutting scheme.

[0040] In an advantageous development, after method step E), the trunk is subjected to a separate processing according to the cutting method, which in particular comprises a substantial separate processing of the trunk, in which, for example, by-products are separated from the trunk after profiling in the trunk.

[0041] As mentioned above, the problem is also solved by an apparatus according to claim 12. The apparatus comprises conveying means provided for conveying the trunks along their trunk axis in a feed direction, and a profile sensor arranged for determining the geometry of the trunks during the conveying movement of the trunks. The apparatus further comprises first separating means arranged and configured to engage the trunks during the conveying movement to form a working surface for separation, an image sensor arranged for inspecting the working surface, and a computing unit configured for determining timber defects and / or timber defect characteristics in the separated processed trunks in a spatially resolved manner and depending on the inspection image and the trunk geometry.

[0042] The device according to the invention is particularly suitable for carrying out the method according to the invention or advantageous developments of the method. In particular, the method according to the invention or advantageous developments of the method can be carried out by the device according to the invention or advantageous developments of the device. In this respect, the above-mentioned remarks about the method apply accordingly with regard to possible embodiments of the device and the advantages that can be achieved thereby. [Brief explanation of the drawings]

[0043] The advantages and possible embodiments of the present invention will be explained below based on exemplary embodiments and drawings. [Figure 1] The trunk geometry of the tree is determined and a virtual trunk model is formed showing the trunk. [Figure 2] The separation and processing of the trunk and the adjustment of the trunk model are shown. [Figure 3] Shows imaging inspection of a tree trunk and projection of the inspection image onto a virtual trunk model. [Figure 4] 13 shows the adjustment of the trunk model according to the inspection image. [Figure 5] Further separation processing according to imaging examination is shown. DETAILED DESCRIPTION OF THE INVENTION

[0044] In industrial timber processing, tree trunks are generally processed into boards derived from different cross-sectional areas of the trunk. Boards derived from the inner cross-sectional area of ​​the trunk are called main-product boards, while boards derived from the outer cross-sectional area are called by-product boards. By-product boards are generally profiled directly on the trunk. In this case, the bark area is separated from the trunk by processing it into blocks or pieces to form a substantially flat work surface. In most cases, the work surface constitutes the long edge of the by-product board. When processing a trunk on multiple sides, the trunk can be a so-called model with work surfaces on two sides after separation, or a block with work surfaces distributed on all four sides. After separation, the separated trunk is scanned by at least one profile sensor. This allows for the detection of internal defects in the timber, such as cracks or branch passages, that have three-dimensional extensions.

[0045] The above-mentioned measures for quality inspection and processing of trunks involve a great deal of metrological effort, but are the result of common practice in timber processing procedures. Based on Figures 1 to 5 below, the method steps are explained, which allow achieving the same advantages with significantly less metrological effort.

[0046] Figure 1 illustrates a tree trunk 1, which has a substantially cylindrical basic geometry extending along its trunk axis 2. The tree trunk 1 has an irregular surface profile due to its natural growth.

[0047] The trunk 1 is conveyed in a feed direction 3 and reaches the detection areas of several profile sensors 4, of which only one profile sensor 4 is shown in the illustrated embodiment. Each profile sensor 4 can be formed, for example, as a laser or light section sensor. The trunk geometry of the trunk 1 is determined as a function of the feed movement, in particular its speed relative to the profile sensors 4, by inputting several profile images into a computing unit 5, which generates a virtual trunk model 6 on the basis of a point cloud. The virtual trunk model 6 has a geometry that substantially corresponds to the geometry of the trunk 1, at least on its outer periphery.

[0048] In the subsequent step shown in Figure 2, the trunk 1 is machined by a rotating cutter head 7 to form a substantially flat work surface 8, the sides of which are bounded by two rounded edges 9. For the sake of clarity, only one of the rounded edges 9 is labeled in Figure 2. The trunk 1, which has been separated here by a method that is likewise not shown in detail, has not only one but, for example, two or, for example, four, distributed work surfaces 8 around its periphery, and after separation can be a so-called model or timber.

[0049] The rounded edge 9 has a characteristic extension associated with the natural growth of the trunk 1 in at least one plane extending substantially along the trunk axis 2 and defined by the working surface 8. The separation of the trunk 1 exposes timber defects 10 at the working surface 8. As mentioned at the beginning, timber defects have a negative effect on the quality of the board to be obtained, the surface of which is at least partly limited by the working surface 8.

[0050] In order to be able to take into account timber defects 10 in the further processing of the trunk 1, a profile cross section is formed on the virtual trunk model 6 in a virtual processing plane 11. Two virtual rounded edges 12 extend along the profile cross section, only one of which is labeled in Figure 2.

[0051] The position and orientation of the virtual processing plane 11 on the virtual trunk model 6 is formed after the trunk separation process according to the nominal rotational position of the trunk 1. This is generally located in the computing unit 5 which controls the separation process of the trunk 1 and its transport movement. The control information determines the nominal rotational position of the trunk 1 during the separation process and makes it possible to position the virtual processing plane 11 relative to the virtual trunk model 6 in such a way that the virtual processing plane extends according to the plane in which the actually formed processing surface 8 of the trunk 1 lies.

[0052] After the step illustrated in Figure 2, the trunk 1 is further conveyed in the feed direction 3 until it reaches the detection area of ​​at least one imaging sensor 13 as illustrated in Figure 3. The trunk is oriented so that the working surface 8 is arranged substantially perpendicular to the measurement axis of the imaging sensor 13. Such an orientation of the trunk 1 can be achieved, for example, by suitable control of the conveying or handling means that move the trunk in the feed direction 3.

[0053] The image sensor 13 detects an inspection image 14 that completely includes the processing surface 8, the timber defects 10 in the processing surface, and the blunt edge 9 of the separated trunk 1. It is also possible to detect only a portion of the processing surface 8 and / or the blunt edge 9 by a method not shown here.

[0054] The computing unit 5 first detects timber defects 10 based on the inspection image 14. This can be done by digital image processing or, for example, by algorithms based on machine learning or AI. The inspection image 14, or only the detected timber defects 10, or an error radar chart corresponding to the timber defects, is then projected into the virtual working plane 11 of the virtual trunk model 6. Importantly, for this purpose, the inspection image 14 is compared with the virtual timber edge 12 based on the timber edge 9 contained in the inspection image, and the inspection image 14, timber defects 10, or error radar chart are positioned on the virtual trunk model 6 based on this. This makes it possible to determine the position and extension of the timber defects 10 and the timber defect characteristics themselves in relation to the trunk geometry of the trunk 1 with great accuracy and spatial resolution.

[0055] Due to unavoidable imprecision in positioning, screwing errors can occur in the trunk 1 during the separation process, which can result in a processed surface being formed on the trunk 1 that is deviated from the desired position and / or orientation of the processed surface. By image inspection of the processed surface 8 on the separated trunk 1, as shown in Figure 4, it is possible to determine the shape or dimensions of the processed surface 8 or, for example, the extent of one or two rounded edges 9 adjacent to the processed surface on the basis of the inspection image 14. It is then possible to determine in what position and / or orientation the virtual processing plane 11 should be in the trunk model 6 in order to obtain an identical or at least similar virtual processed surface 15 or virtual rounded edges 12 according to the profile cross-section of the trunk model 6.

[0056] As shown exemplarily in Fig. 4, the trunk model 6 can be repeatedly cut, for example, in a number of possible cutting planes, and a number of corresponding virtual machining surfaces 15 or virtual rounded edges 12 can be determined on the trunk model 6 based on the respective cuts. The computing unit 5 compares the detected inspection image 14, and in particular the machining surface 8 and / or rounded edge 9 contained in the inspection image, with the cut of the trunk model 6 and the virtual machining surface 15 and virtual rounded edge 12 contained in the cut. If there is sufficient similarity, the corresponding cut plane is defined as a virtual machining plane 11 and the inspection image 14 or, for example, an error radar chart is projected into the cut plane. It is also conceivable to input the machining surface and / or rounded edge extension as well as the trunk model 6 into a computing model and determine the position and / or orientation of the virtual machining plane 11 explicitly, in particular in one computing step, based on the similarity criterion.

[0057] By accurately and spatially resolving the detection of timber defects 10 or timber defect characteristics, it is possible to determine cutting strategies that allow the trunk 1 to be processed for optimal timber yield. This is illustrated in FIG. 5. For example, a by-product can be profiled by milling two adjacent rounded edge areas 16 of a replica, or the division of two adjacent by-product boards can be altered by a split plane 17, so that only one by-product board is affected by the timber defect 10 and the other by-product board is not affected.

Claims

1. A method for detecting timber defects and / or timber defect characteristics when processing tree trunks, comprising the method steps of: A) performing a profile measurement of the tree trunk (1) and determining the shape and dimensions of the trunk; B) forming a processing surface (8) along the trunk axis of the trunk (1) and then performing separation processing of the trunk (1); C) performing an imaging inspection of the separated and processed tree trunk (1) at least at the processing surface (8) to obtain at least one inspection image (14); D) A step of spatially resolving determination of timber defects (10) and / or timber defect characteristics in the tree trunks separated and processed according to the inspection image (14) according to method step C) and the trunk geometry according to method step A).

2. 2. A method according to claim 1, characterized in that the spatially resolved determination of the defects in the timber and / or the defect characteristics of the timber is carried out by generating a virtual trunk model (6) based on the geometry of the trunk according to method step A), generating at least one profile cross-section of the virtual trunk model (6) in a virtual working plane (11), and projecting the inspection image (14) acquired in method step C) or information derived from the inspection image onto the virtual trunk model (6) in the virtual working plane.

3. Said method step D) comprises two substeps D1) and D2), detecting in said sub-step D1) said defects (10) of the timber and / or said defect characteristics of the timber in said inspection image (14), 3. The method according to claim 2, characterized in that in substep D2) the position and / or extent of the timber defects and / or timber defect characteristics in the separated and processed tree trunk (1) are determined according to common features between the inspection image (14) and the virtual trunk model (6).

4. 3. A method according to at least claim 2, characterized in that by the profile cross-section in the virtual working plane (11), at least one virtual round edge (12), in particular two virtual round edges (12), are formed in the trunk model (6), and a spatially resolved determination of the timber defects (10) and / or the defect characteristics of the timber is carried out depending on the round edges.

5. performing said imaging inspection in said method step B) to detect a rounded edge (9) adjacent to said work surface (8); 5. The method according to claim 4, characterized in that in method step D) the spatially resolved determination of the defects and / or the defect characteristics of the timber is based on a comparison of the virtual round edge (12) of the trunk model (6) with the round edge (9) adjacent to the working surface (8) of the separately processed trunk (1).

6. 3. The method according to at least claim 2, characterized in that the position and / or orientation of the virtual working plane (11) on the virtual trunk model (6) is determined in method step B) depending on the nominal rotational position of the tree trunk (1).

7. 3. The method according to at least claim 2, characterized in that the position and / or orientation of the virtual working plane (11) in the virtual trunk model (6) is determined depending on the working surface (8) and / or at least one rounded edge (9) of the inspection image (14).

8. 8. The method according to claim 7, characterized in that the virtual working plane (11) is determined in the virtual trunk model (6) by determining a virtual working surface (15) and / or a virtual rounded edge (12) that is identical or at least similar to the working surface (8) and / or the rounded edge (9) of the inspection image (14), and the position and / or orientation of the virtual working plane (11) is selected in particular so that the virtual working plane includes the determined virtual working surface (15) and / or virtual rounded edge (12).

9. 8. The method according to claim 7, characterized in that the method step D) is carried out independently of the dimensional and geometrical information of the processed surface (8) of the separated trunk (1).

10. 10. The method according to claim 1, further comprising determining the cutting strategy as a function of the spatially resolved defects (10) in the timber and / or the defect characteristics of the timber in method step E).

11. 11. The method according to claim 10, characterized in that after method step E), the trunk (1) is separated and processed according to the cutting method.

12. 12. An apparatus for processing tree trunks, in particular for processing tree trunks by a method according to any one of claims 1 to 11, comprising: a conveying means for conveying the tree trunk (1) along its trunk axis in a feeding direction (3); a profile sensor (4) arranged to determine the geometry of the trunk (1) during its transport movement; a first separating means (7) arranged and configured to engage said tree trunks (1) to form a working surface for separating them during said conveying movement; an image sensor (13) arranged to inspect the work surface (8) and output an inspection image (14); and a computing unit (5) configured to determine timber defects (10) and / or timber defect characteristics in the separated processed tree trunk in a spatially resolved manner and depending on the inspection image (14) and the shape of the trunk.

Images