A device for inspecting an object in a moving flow and a machine including such a device

Abstract

The present invention relates to an apparatus (1) for inspecting an object (2) moving in a flow (F) on a surface (3) moving in a direction (D), wherein the apparatus (1) comprises an illuminated strip (6) of the surface (3), a detection means (7) having an optical axis (AO), and finally, an illumination means (4) that processes and evaluates a signal supplied by the detection means (7) to detect the presence of the moving object (2) and map its height or / or determine its volume. The device (1) is characterized in that the median plane (PM) of the light ray (5) and the optical axis (AO) of the detection means (7) have a predetermined angle (AP) between them, the image of the light strip (6) obtained by the detection means (7) has a width equal to at least three times the resolution of the means (7), and the light means (4) is configured such that the light strip (6) generated on the support surface (3) is divided by two opposing straight parallel edges and has sufficient contrast for the non-illuminated area (6') regardless of the color of the object (2).

Description

Technology Field

[0001] The present invention relates to the field of inspecting moving objects, and in particular to the field of a machine that automatically sorts and inspects objects moving in a fluid manner on, for example, a conveyor belt type conveyor. Generally, this type of machine is used to sort household waste, and the conveyor used in this situation generally has a width of 1 to 3 m and the height of the moving object is up to about 300 mm.

[0002] The technical concept of the present invention is, advantageously but not exclusively, to an inspection device that can be implemented in the aforementioned field, as well as a machine comprising at least one such device and a method for carrying out said device, in particular, for detecting the presence and / or dimensions or geometric features of an object that cannot be recognized by conventional inspection techniques using reflection or radiation. Background Technology

[0003] In fact, in the field of classification and inspection involving the application of radiation, some objects cannot be recognized by detectors because they lack any spectral shape or features that would allow them to be distinguished on the conveyor belts in which they are placed. In the near-infrared spectral region (below 1,000 nm to 2,500 nm), this is the case for black objects, such as carbon black, metals, or inert materials (e.g., glass, earth, or stone). In the visible wavelength range (between 400 nm and 800 nm), objects that cannot be recognized by color vision systems are those that do not exhibit contrast or exhibit very low contrast with their surroundings. For example, on a black conveyor belt, black or transparent and colorless objects are difficult or even impossible to detect.

[0004] However, for reasons of cost, durability, and reliability, most conveyors in sorting machines are black. More specifically, they are black when new, but because they are generally not cleaned, they turn dark gray or brown due to dust after use.

[0005] Accordingly, detecting these moving objects and, where appropriate, characterizing them is proposed by mapping all material points located on the support plane of the conveyor, preferably by determining their height relative to this plane, that is, by mapping all points of all objects on the said plane of the conveyor.

[0006] Currently, several different solutions used industrially for this purpose are as follows:

[0007] time-of-flight measurement;

[0008] stereovision cameras;

[0009] laser profilometer

[0010] The first solution involves emitting radiation or ultrasound from a source located at a predetermined height and measuring the return time of the waves to the emitter, specifically measuring the return time after the waves have rebounded from the measured object. Its precision is directly related to the performance of focusing the emitted waves to an accurate point and the precision of measuring the return time. Focusing constraints orient this solution to use laser light and scanners, and the use of light requires electronic devices capable of measuring picoseconds to obtain a resolution of the order of 1 mm. Therefore, this solution requires highly specialized, very expensive, and delicate electronic devices.

[0011] The other two known methods mentioned earlier are purely geometric.

[0012] In stereovision, parallax is used between two cameras spaced 20 to 30 cm apart and facing substantially the same direction. In a manner very similar to human vision, the lateral offset of the same point seen by the two cameras is directly linked to the distance of that point to each camera. This solution requires two cameras and typically controlled lighting. While object analysis is performed by image processing software to establish a correspondence between the locations of feature points of an object in the two images, it is often very useful, or even necessary, to add structured lighting, such as lasers, to clearly recognize and locate the corresponding points between the two images.

[0013] When using a laser profilometer, the parallax between the illumination direction and the detection direction is also utilized. A single camera is required, and therefore this solution is substantially simpler than stereo vision.

[0014] Nevertheless, a common risk of these two geometric solutions is related to the effects of shading and masking: for good detection, the object point under consideration must be in the direct view (obstacle-free line of sight) of the light source and one or more associated cameras. Obviously, this condition is better met with two lines of sight (laser profilometers) than with three (stereo vision cameras).

[0015] Consequently, laser profilometers are preferred and are often employed in inspection and classification fields. It is also understood that there is a compromise between the risk of masking and the precision of height measurement, as both increase with parallax.

[0016] For example, documents WO 2012 / 089185 and WO 2019 / 043231 disclose the use of a laser profilometer to measure the height profile of an object moving on a mat. These documents utilize the principle that if the illumination is sufficiently strong, even black or transparent objects return a significant and detectable signal. (In contrast, clearly opaque objects return a very strong signal, saturating the sensor.) Due to its typical beam diameter of about 1 mm, a collimated laser naturally provides this illumination.

[0017] Document WO 2012 / 089185 specifically allows only coarse processing to detect the extension of an illuminated strip associated with such an optical path, using the diffused laser light into the material and the resurgence of the material after it passes beneath the surface of the material, along with the generation of images of an illuminated strip with fuzzy limits and variable width. It should be noted that this operation requires the use of collimated lighting (and thus laser) to obtain an illuminated strip of constant width regardless of the height of the illuminated surface.

[0018] A known disadvantage of using lasers is that protective and optical precautions must be provided to avoid the risk of glare and eye damage to workers moving near the machine. Therefore, the machine must be carefully covered to prevent any direct rebound of the laser beam outside the machine. However, when sorting heterogeneous and diverse types of objects, such as household waste—whether packaging or of industrial origin—numerous metal surfaces exist, including cans, metallized films, exposed external or internal metal coatings, and various metal parts. These metal surfaces act as almost perfect mirrors, with the exception of painted surfaces.

[0019] However, the flow path of a moving object must generally maintain a minimum height of 30 cm without any obstacles to avoid any risk of jamming. Due to these two combined constraints, it is almost impossible to guarantee that the laser beam will not rebound to external points.

[0020] This problem is very similar to the problem occurring in X-ray machines where the solutions employed to address radiation use tunnels several meters long and even baffles at the machine entrance to circulate the object. Such solutions are clearly not applicable to the mentioned situation.

[0021] One solution sometimes proposed for this problem involves using infrared (IR) lasers, which are considered safe for the human eye because they operate outside the range of human senses. However, the risk of burns remains, and this is even more dangerous because IR lasers are invisible, leading people to pay little attention and tend not to adopt protective measures. The problem to be solved

[0022] The objective of the present invention is to propose a solution for overcoming the disadvantages of known inspection means based on laser profilometers, while providing similar performance capabilities in terms of detection and inspection for moving objects, particularly black and dark objects. means of solving the problem

[0023] To this end, the first object of the present invention is an apparatus for inspecting an object moving in a substantially monolayer flow on a flat support surface moving in a longitudinal direction, the apparatus comprising at least one lighting means for providing a lighting beam having a median plane or axis that extends across the direction of movement and across all or part of the width of the support surface, wherein the beam is directed toward the support surface on one side and the intersection with the support surface is composed of a lighting strip; On the other hand, at least one detection means having an optical axis or plane and an acquisition field, the dimensions of which are at most equal to the dimensions of the light strip in a direction transversing the direction of movement, and also configured to encompass a volume extending above the non-illuminated area of ​​a support surface located upstream and downstream of the light strip in the direction of movement and above at least the light strip or a portion of the strip, and optionally above the non-illuminated area; and finally, means for processing and evaluating a signal or data provided by the at least one detection means, in a view for detecting the presence of the moving objects, and in a view for mapping their height and / or determining their external volume, wherein the objects are preferably dark or black, and the intermediate plane or axis of one or more light beam(s) and the optical axis or plane of the detection means have a non-zero angle between them called a parallax angle, and the light strip acquired by the at least one detection means The image has a width equal to at least three times, preferably at least ten times, the resolution of the above means.

[0024] The above device is characterized in that the at least one lighting means is configured and arranged such that the lighting strip (6) created on the support surface has a width greater than 10 mm in the direction of movement, advantageously between 10 mm and 30 mm, is delimited by two opposing straight parallel edges that form a clear limit in the image of the lighting strip when there is no object, and has sufficient contrast for the non-illuminated area so that it can be detected by the detection means and the processing and evaluation means regardless of the color of the moving object to which the lighting strip is applied, and one or each lighting beam is composed of a beam derived from an incoherent light source.

[0025] The present invention also relates to an inspection and sorting machine comprising at least one such device and a method for inspecting an object moving in a fluid manner on a support surface. Brief explanation of the drawing

[0026] The present invention will be better understood by the following description regarding preferred embodiments, which are provided by non-limiting embodiments and are described with reference to the accompanying schematic drawings: [Fig. 1a] Side view of a preferred embodiment of an inspection device according to the present invention, wherein the detection means (camera) is located upstream and low to the illumination; [Fig. 1b] A side view similar in form to Fig. 1a, but the detection means (camera) is positioned upstream and high above the lighting means, and is maintained under the lighting; [Fig. 2] This is a perspective view of the device shown in Fig. 1a; [FI. 3] A perspective view of another preferred embodiment of the device according to the present invention, wherein the detection means (camera) is located downstream of the illumination means and faces it (illumination means) (opposite side of a plane transverse and perpendicular to the support surface); [Fig. 4] is a perspective view of another preferred embodiment of a device according to the present invention using a movable lighting means; [Fig. 5] is a perspective view showing the intersection of a rectangular object moving on the support plane and the lighting strip; [Fig. 6] is a view from the plane of the camera showing how the image of the light strip is modified by the presence of the object of Fig. 5; [Fig. 7] shows the height profile inferred from the information in Fig. 6; [Figs. 8a], [Figs. 8b], [Figs. 8c], [Figs. 8d], [Figs. 8e] and [Figs. 8f] show pairs of [momentary images of light strips / accumulated height profiles] when the entire object is circulated through the device according to the present invention. Specific details for implementing the invention

[0027] FIGS. 1 to 4 schematically illustrate various embodiments of a device for inspecting an object (2) moving in a substantially monolayer flow (F) on a flat support surface (3) moving in the longitudinal direction (D).

[0028] This device (1) basically includes the following.

[0029] On the one hand, at least one lighting means (4) providing a lighting beam (5) having a median plane or axis (PM) that extends across the direction of movement (D) and across all or part of the width of the support surface (3), and a lighting strip (6) that is directed toward the support surface (3) and has an intersection with the support surface (3).

[0030] On the other hand, at least one detection means (7) having an optical axis or plane (AO) and an acquisition field (7'), the dimensions of which are at most equal to the dimensions of the lighting strip (6) in a direction across the direction of movement (D), and also configured to encompass a volume extending above the lighting strip (6') and at least above the lighting strip (6) and optionally above the non-illuminated area (6') of a support surface (3) located upstream and downstream of the lighting strip or part (6) in the direction of movement (D), and also above the non-illuminated area (6').

[0031] And finally, means (8) for processing and evaluating the signal or data provided by the at least one detection means (7) in view of detecting the presence of the moving objects (2), and in view of mapping their height and / or determining their external volume.

[0032] The axis or intermediate plane (PM) of one or more illumination beam(s) (5) and the optical axis or plane (AO) of one or more detection means(s) (7) have a non-zero angle (AP) between them called the parallax angle. Additionally, the image of the illumination strip (6) obtained by the at least one detection means (7) has a width equal to at least three times the resolution of the means (7), and preferably at least ten times the resolution.

[0033] The term “n times the resolution” is understood to mean a width value corresponding to n times the dimension of a component or basic detection unit in this specification. Accordingly, for example, in the case of a matrix detection means such as a camera or similar image acquisition means, the width of the image of the strip (6) will be at least 3 pixels (a strip formed by at least 3 rows of pixels).

[0034] In any case, in practice, the width of the lighting strip (6) on the support surface (3), that is, the distance separating two adjacent edges (9, 9'), must be sufficient to easily differentiate and clearly distinguish these two edges so as to make advantageous use of the information provided individually by each of them. The inventors were able to determine, by various tests, that the width of the lighting strip (6) formed on the movable support surface (3) should be greater than 10 mm in the direction of movement (D), advantageously between 10 mm and 30 mm, preferably on the order of about 20 mm. The wider the width of the part, the more expanded lighting power is required to have sufficient contrast.

[0035] Accordingly, according to the present invention, the at least one lighting means (4) is configured and arranged to have sufficient contrast with respect to a non-illuminated area (6') so that when the lighting strip (6) created on the support surface (3) has a width in the direction of movement (D) greater than 10 mm, advantageously in the range between 10 mm and 30 mm, and there is no object (2), it is defined by two opposing straight parallel edges (9, 9') that form a clear limit in the image of the lighting strip (6), and can be recognized (detected) by the detection means (7) and the processing and evaluation means (8) regardless of the color of the moving object (2) to which the lighting strip (6) is applied. Additionally, one or each lighting beam (5) is composed of a beam originating from a non-coherent light source (4'). Therefore, by the configuration of the present invention, it is possible to overcome the disadvantages and risks associated with implementing a laser profilometer, and at the same time, similar performance capabilities can be provided in terms of detecting and mapping moving objects (2), and especially black, dark, and transparent objects.

[0036] The concept of a dark or black object can be characterized by luminosity below a reflectance threshold, expressed as a percentage of the reflectance of a white reference object (e.g., white ceramic or white office paper). Typical values ​​are reflectances between 2% and 10% for objects perceived as black by the human eye.

[0037] As described above, the lateral delimitation of the light strip (illuminated strip, illuminated strip, illuminated strip) (6) and its light intensity are advantageous such that the area it limits has two clear linear limits (9, 9'), and beyond them the light intensity drops abruptly by, for example, at least 50%, advantageously at least 60% to 80%. Such contrast by factors of 2 to 5 is sufficient to allow a clear distinction between the illuminated object (2), even dark or black, and the unilluminated part of the support surface (conveyor) (generally black, brown, or dark gray).

[0038] When moving upward from the above conveyor (at a predetermined distance from the support surface) and thus from the focus (transverse focusing line), the lighting area widens: however, the limit clearly remains and is clearly identifiable in the image of the lighting strip (6). Thus, the device (1) maintains efficiency up to a height of approximately 10 cm, which is generally more than sufficient for presence detection and even height measurement, because 99% of the objects (2) circulating on the conveyor are less than 10 cm in height in the situation of a preferred application of the present invention (processing of household and industrial waste).

[0039] A person skilled in the art can likely consider this to be a type of structured lighting having two detection lines (9 and 9') instead of a single line in the case of a laser: the two "lines" are formed by linear boundaries or edges between a highly illuminated area (illumination strip) and the outside (particularly the strip immediately adjacent laterally and unilluminated). Thus, the information provided is more substantial than the information resulting from the use of a reflected laser beam (twice the collected information).

[0040] A person skilled in the art also understands that, depending on the width of the movable support surface (3) and the dimensional / performance capability features of the means (4 and 7), the device (1) may include one or more lighting means(s) (4) and / or one or more detection means (7) arranged side by side so that the device (1) can perceive the entire width for the purpose of inspecting the object (2). In any case, the acquisition field of one or each acquisition means (7) is completely contained within the range of the lighting strip (6) in the direction (D) as described above.

[0041] Advantageously, the parallax angle (AP) is greater than 10°, preferably in the range between 10° and 50°, and the lighting means (4) and the detection means (7) may or may not be located on the same side of a plane (PP) perpendicular to the support surface (3) and parallel to the lighting strip (6) (see FIG. 1 to 4).

[0042] Of course, the detection means (7) may be composed of several or even many components, and its signals are integrated before being used by the means (8).

[0043] However, advantageously, the device (1) includes a single detection means (7).

[0044] One or each detection means (7) is preferably composed of a matrix camera, and the representation of the light strip (6) used in the detection means (7) and / or the processing and evaluation means (8) is advantageously composed of a line (LP), which preferably extends from the bottom of the image provided by the detection means (7) and analyzed by the processing and evaluation means (8), as shown in FIG. 7, in the case where there is no object (2) impacted by the light strip (6).

[0045] Preferably, one or each lighting means (4) comprises at least one non-coherent light source (4') arranged across the direction of movement (D) having a constitution, and even a plurality of such sources arranged across. Additionally, it comprises at least one element (4'') that focuses (concentrates) the radiation emitted by the light source (4') to create a lighting strip (6) on the support surface (3).

[0046] According to the first configuration variation shown in FIGS. 1 to 3, one or each lighting means (4) provides permanent lighting on the strip (6) and includes at least one light source (4') in the form of a straight filament extending across the direction of movement (D) (e.g., a halogen lamp) and a focusing element (4'') in the form of a reflector having a partially cylindrical-elliptical cross section, wherein the filament (4') is positioned at one of the focal points of the virtual ellipse of the reflector (4'') and the other focal point is located near the support surface (3).

[0047] According to the second configuration variation shown in FIG. 4, one or each lighting means (4) provides lighting scanning over a lighting strip (6) and includes a fixed light source (4') connected to a focusing element (4''), for example, in the form of a filament, for example, in the form of a reflector having a partially elliptical cross section, and the lighting means (4) further includes a movable transmission element (4''') to create a lighting strip (6) by applying lighting originating from a source (4') and focused on a support surface (3), for example, by transverse scanning.

[0048] Creating a lighting strip (6) that exhibits sufficient geometric and luminous characteristics for reliable operation of the device (1) requires considering specific points and references for the purpose of obtaining straight and straight edges (9, 9') on the strip, particularly in relation to focusing non-coherent and multi-directional radiation originating from the source (4') within the context of configuring and mounting the components of one or each lighting means (4). The latter, or rather the latter image, may theoretically be limited in width direction to a minimum dimension corresponding to three columns of pixels in an image acquired by the camera (7), but in practice, it has a width of at least 10 columns of pixels, physically has a width of one or more tens of millimeters, and preferably has a width of at least 20 mm.

[0049] Therefore, regarding the problem of focusing blur, it should be noted that the lighting strip (6) is an enlarged image of the filament of the source lamp (4') located at the first focus of the virtual ellipse of the associated cylindrical-elliptical reflector (4'') in the dimensions of the lighting strip parallel to the direction of travel (D) of the conveyor. If this filament is off-center relative to the focus, the image becomes blurry.

[0050] In addition, for this type of lamp, the filaments visible across the conveyor take the form of a daisy chain (5 to 7 per filament) supported by supports. This shape bends and moves away from the focal point, especially when it is heated and thus expanded. Consequently, the image produced by the various parts of the filament is substantially blurred.

[0051] Finally, it should also be noted that the two transverse edges of the one or each of the reflectors (4) are flat sheets. These are positioned perpendicular to the light strip (6) and return the ray received from the filament (4') to the strip. Therefore, a person skilled in the art will understand that special attention needs to be paid to the design (rigid parts) and its perfectly perpendicular attachment.

[0052] Due to the structure of the elliptical or cylindrical-elliptical reflector (4), the illumination beam (5) emitted by the means (4) consists of two distinct cumulative parts that provide two mutually overlapping illumination areas on the support surface (3):

[0053] - An illumination area focused on the direction of progress or movement (D) of the object (2) exhibiting high density: This is provided by a beam passing through a cylindrical-elliptical reflector (4'') and corresponds to an enlarged image of the filament of the halogen tube used. In a direction traversing a conveyor or belt corresponding to a movable flat support surface (3), the beam is positioned over the entire width;

[0054] - A direct lighting area around the preceding zone connected to the light rays reaching the direct line of the filament of the halogen lamp (4') used. The maximum density of this part is less than 1% of the density of the previous one. By adjusting the sensitivity of the camera (7) appropriately, this lighting component can be overlooked compared to the preceding focused part.

[0055] The illumination means (4) that can be used within the scope of the present invention has already been mentioned in literature EP 1243350 and EP 3423202 (near-infrared spectroscopy) in relation to the use of the spectral composition of reflected radiation to classify materials. Accordingly, the inspection device (1) according to the present invention can be integrated at a lower cost into the type of machine described in these literature and already commercially available by the applicant, because the illumination means (4) can be common to two separate detection means in which it is implemented, and advantageously, the collected signal can be used by the same processing and evaluation means (8).

[0056] Another objective of the present invention is a method for inspecting an object (2) moving in a substantially monolayer flow (F) on a flat support surface (3) moving in the longitudinal direction (D).

[0057] This method is carried out by implementing a device (1), said device comprising, on the one hand, at least one illumination means (4) providing an illumination beam (5) having a median plane or axis (PM) that extends across the direction of movement (D) and substantially across the entire width of said support surface (3), and on the other hand, at least one detection means (7) having an optical axis or plane (AO) and an acquisition field (7'), said detection means (7) configured to encompass at least the illumination strip (6), a non-illuminated area (6') of said support surface (3) located on either side along the illumination strip (6), and at least one volume extending above the illumination strip (6) and optionally above the non-illuminated area (6'), and finally, in view for detecting the presence of said moving objects (2), and their height It includes means (8) for processing and evaluating signals or data provided by at least one detection means (7) in terms of mapping and / or determining their external volumes.

[0058] The device (1) is configured such that an optical axis or plane (AO) and an intermediate plane or axis (PM) of the illumination beam (5) have a non-zero angle (AP) between them called a parallax angle, and the method comprises, on the one hand, an image of the illumination strip (6) obtained by the at least one detection means (7) having a width at least equal to three times the resolution of the means (7), and on the other hand, an illumination strip (6) on a support surface (3) that is bounded by two opposing straight parallel edges (9, 9') that are mutually distinguishable in the image, having sufficient contrast for a non-illuminated area (6') so that it can be used by the processing and evaluation means (8) and detected by the detection means (7) regardless of the color of the moving object (2) to which the illumination strip (6) is applied, and the width (D) in the direction of movement is greater than 10 mm and advantageously between 10 mm and 30 mm, wherein the illumination beam (5) or each of the illumination beams originates from a non-coherent light source (4'). It consists of beams.

[0059] According to one feature of the present invention, at least one profile line (LP) representing the elevation of a moving support surface (3) is extracted from an image of a lighting strip (6) by at least one suitable data processing program executed by a processing and evaluation means (8), and this profile line is compared with a minimum threshold value for the presence of an object, and, where appropriate, this profile line is used to determine the height of the object (2) to be inspected point-by-point.

[0060] According to another feature, it may also be possible to analyze the luminance of the continuous profile lines (LP) of the object (2) and use this to classify the object (2) as a dark object or a clear object.

[0061] Finally, it may also be possible to accumulate consecutive profile lines of the same object (2) over time to reconstruct the apparent shape and / or to compute its volume.

[0062] Advantageously, the profile detection program applied by the processing and evaluation means (8) can perform a search for the luminance percentiles detected along each line perpendicular to the image of the lighting strip (6), and then perform an operation (operation) to threshold the profile lines.

[0063] In relation to an alternative embodiment, the processing and evaluation means (8) may perform two profile detection operations (operations) to find and use two limits, namely the high and low intersections of the existing object (2) and the light strip (6), that is, two edges (9 and 9') of the image of the light strip (6).

[0064] Accordingly, the above method may include determining a first profile following an upper limit (LH) defined by a first angle (α), such as a first edge of a light strip in the direction of movement, for example, and a second profile following a lower limit (LB) defined by a second angle (α2), such as a second edge of a light strip in the direction of movement, for example, and these two limits (LH, LB) correspond to the two aforementioned edges (9, 9') of the light strip (6) projected onto the object (2) and are offset in time according to the movement of the object under inspection (2).

[0065] Preferably, the two aforementioned operations (operations) are implemented separately and individually, and the two profiles are determined independently of each other.

[0066] Also, preferably, the above-described method implements the inspection device (1) as described above.

[0067] Finally, the present invention also relates to a machine for automatically inspecting and sorting an object (2) moving in a substantially single-layer flow (F) on a conveyor belt type conveyor, for example, which provides a moving support surface (3) for the object (2), wherein the machine is characterized by including at least one inspection device (1) as described above, which is mounted on the conveyor.

[0068] Optionally, several (at least two) devices (1) may be provided, each comprising a common lighting means (4) and at least two separate detection means (7), or a common detection means (7) and at least two separate lighting means (4), or even each comprising a lighting means (4) and a detection means (7) and corresponding to at least two separate modules arranged side by side across the direction of movement (D).

[0069] A more detailed description of various alternative embodiments of the present invention is provided below with reference to the attached drawings.

[0070] First, referring to FIG. 1a, the lighting means (4) is illustrated as a lighting fitting (4) having a source (4'), namely a focused reflector (4'') and a halogen tube (4'), hereinafter referred to as a "halogen reflector," which concentrates light energy onto a narrow lighting strip (6) on the surface of a conveyor (3) transporting an object (2) to be inspected. The light beam arrives at an average angle α with respect to the vertical portion of the moving flat transport support surface (3), and the beam (5) is limited by extreme angles α1 and α2. In this case, a matrix camera (7) is positioned upstream of the lighting area in the form of a strip (6) and is directed toward the conveyor at an average angle β (of its optical axis (AO)) with respect to the horizontal (flat surface of the conveyor). The bottom and top of its field of view or acquisition volume (VA) correspond to angles β1 and β2.

[0071] Compared to a laser profilometer, the main difference is that the illumination is not only focused in direction (α), but also fills the entire angular sector from α1 to α2. The width of this illumination strip (6) depends on the variable height (above the conveyor) of its application surface (conveyor surface or surface of an object), but its two limits (9 and 9') can be clearly detected by the camera (7) in the aforementioned height range and remain sharp.

[0072] This light strip (6), illuminated by focused light, is functionally combined with a matrix camera (7) to form a triangulation system capable of detecting and mapping all objects (2) present on the conveyor and passing through the light strip, even black, dark, and spectrally inert objects.

[0073] These devices (1) have several advantages in terms of cost and integration: one camera per device, no specific lighting and no covering costs (because there is no laser source).

[0074] Finally, although not illustrated in this specification, the machine may include, in addition to the detection means (7) of the device (1), another detector positioned on the direction detection plane α between the two reflectors (4'') to form a detection system that is coplanar with the illumination and inclined in the vertical section, as described in the aforementioned document EP 1243350.

[0075] The mean parallax of the device (1) is given by the difference between two lines of sight, or in this case, the angle AP = π / 2 - α - β. The larger this angle, the more accurate the height measurement, but the greater the risk of masking depending on the shape of the apparent object (2).

[0076] Advantageously, the value selected for α is in the range of 0 to 30°, and the value selected for β is in the range of 30 to 75°.

[0077] Figure 1a shows a value β close to 30° ("low upstream" position), and Figure 1b shows a "high upstream" position where β close to 75°.

[0078] Some other preferred embodiments of the device (1) use different geometric arrangements without departing from the scope of the invention.

[0079] Therefore, two possible positions can be considered for the camera and are illustrated as perspective views in FIGS. 2 and FIGS. 3.

[0080] FIG. 2 shows a "low upstream" position (a configuration where the angle between the camera's axis of sight and the vertical part is important). This position is also illustrated in FIG. 1a. In this case, the camera's degree of freedom of movement in the horizontal direction can be maintained, and this variation demonstrates excellent accuracy for altitude measurement. However, it can also produce an object masking effect (upstream objects can obscure objects on the illuminated strip (6)).

[0081] FIG. 1b illustrates an example of a "high upstream" position: in this case, the camera (7) is installed as close as possible directly below the lighting means (4). Thus, parallax is minimized, which has the advantage of reducing the masking effect compared to the aforementioned configuration, but reduces altitude accuracy compared to the aforementioned configuration.

[0082] FIGS. 1 and 2 show that the lighting means consists of two halogen reflectors (4' + 4'') spaced apart from each other in the direction across. This possible configuration is not strictly necessary, and the operation of the device (1) would be the same as a contiguous row of reflectors (4' + 4'') covering the entire width of the conveyor (3). In practice, if the parallax (angle AP) is sufficient, the field of view (VA) of the camera (7) is located entirely below the lower limit of the reflectors (4) and does not interfere with the reflectors.

[0083] FIG. 3 illustrates an example of a position called a “downstream” position corresponding to a high configuration looking backward (opposite to the direction of movement (D) of the flow (F)) with the possible degrees of freedom of the camera (7) in the vertical direction. This configurational variation experiences reduced masking effects between objects (2) because the objects begin to fall from the end of the conveyor, but experiences greater exposure to soiling from the waste discharge system located just below the end of the conveyor.

[0084] Finally, FIG. 4 illustrates a configuration using a movable lighting means (4) having the same features as the fixed lighting means of FIG. 1 to 3. This type of lighting is described, for example, in document WO 2013 / 115650.

[0085] As illustrated in FIG. 4, radiation from a source (4') including an incandescent lamp, e.g., a halogen lamp, is focused by a convergent reflector (4'') or by a lens and projected onto a rotating mirror (4'''), which always sends the focused light to an elementary rectangular zone, which scans a light strip (6) on a moving flat support surface (3). This strip is naturally separated by the limits of the mirror surface that return the light beam of the lamp (4'). For the correct operation of the device (1) of FIG. 4, it is desirable to have maximum contrast at the edges (9 and 9'). This condition is satisfied when the convergent beam (5) originating from the lamp (6) intersects the upper and lower edges of the mirror surface. Thus, two clear limits of the illuminated strip are obtained, and the same applies even when they are substantially further apart than those of the strips obtained by the variations of FIGS. 1 to 3, as in the solution described above.

[0086] When there is no object (2), the light strip (6) appears as a horizontal strip in the image of the camera (7). The camera's field of view (VA) must be adjusted so that this strip is visible in the lower half of the image.

[0087] Furthermore, and in a manner known to a person skilled in the art, the exposure time and / or diaphragm of the camera (7) are adjusted so as not to saturate the light line and the matte black object. A filter may also be added to the objective lens to adjust the light intensity range to the camera's sensitivity dynamics.

[0088] When the object (2) passes, the position of the depiction of the light strip (6) in the image is modified as shown in FIG. 5 (direction Y indicates the direction of progress or movement (D) of the moving support surface of the conveyor (3).

[0089] In the measured image, the useful collected data is in the form shown in Fig. 6.

[0090] The first profile detection program (e.g., using an image segmentation technique) searches for the location of the maximum signal column-by-column. This maximum value is typically located in the intersection area of ​​the light strip (6) and the object (2).

[0091] To make this process more stable, and as mentioned above, this search for the maximum value can be replaced with a column-by-column percentile search. In practice, the maximum value is at risk of being obstructed by various phenomena (dead pixels in the camera sensor, bright spots in the image, etc.). Since the illuminated strip (6) is relatively wide (e.g., compared to a laser beam) and measurable across at least several wide pixels in the image, the use of high percentiles allows this type of measurement bias to be overcome. The targeted value for this percentile is between 80% and 100% (100% corresponds to the maximum value). Thus, the profile line computed by this first program corresponds to a vector containing as many elements as there are columns in the image. The value of this vector is expressed in pixel numbers, and simple geometric modeling shows that this number is substantially proportional to the height of the object relative to the level of the support surface of the conveyor (3). Then, thresholding of this profile line is performed to isolate the object (2) from the rest of the support surface (3). Once processed, the profile signal (profile line LP) takes the shape shown in FIG. 7.

[0092] In this diagram, the solid line represents the profile line measured in the image, and the dashed line represents the location of the threshold at which the object (2) is considered to have been detected.

[0093] A triangulation calculation known to a person skilled in the art determines the height of the object (2) at this point from the angle of FIG. 1a and the distance between the camera (7), the lighting means (4), and the supporting surface of the conveyor (3). This height is substantially proportional to the vertical position of the strip represented by pixels and is calculated from the position of the strip when there is no object (zero line).

[0094] The minimum detectable height theoretically corresponds to an offset of at least one pixel relative to the zero line, and a reliable minimum value is obtained for two offset pixels. In a preferred embodiment, this is 1 to 3 mm as a function of the resolution of the detection means (7) / camera used.

[0095] All objects (2) having an altitude above a predefined threshold can be distinguished by the result of an operation performed by the aforementioned computation program implemented by the processing and evaluation means (8).

[0096] Therefore, the first service provided by the device (1) is to detect the presence of any object (2) having a minimum height.

[0097] The implementation of a second profile detection program may be provided (preferably in addition to the first program), which is intended to detect two limits, namely the high and low intersections of the existing object (2) and the light strip. This additional information provides a certain amount of redundancy and can therefore be used to provide a more accurate diagnosis of the shape of the object under inspection (2). For example, in FIG. 1a, the profile can be equally determined by following the upper limit LH defined by angle α1 and the lower limit LB defined by angle α2, and these two limits (upper limit, lower limit) correspond to the edges (9, 9') of the light strip (6) projected onto the object (2). Thus, two profiles are obtained that are offset according to time as the object under inspection (2) moves. Having two independently calculated profiles increases the average accuracy. Additionally, the entry and exit times of the object (2) at each limit are different. In this case, the vertical part of the object (the edge of the cardboard box) can be measured more easily, or any masking effect caused by an object located upstream of the object of interest can be avoided.

[0098] As can be seen from Fig. 4, in addition, this dual detection of the two lower and upper limits of the lighting strip (illuminated strip) can also be performed at the same angles α1 and α2.

[0099] Additionally, it is possible to consider distinguishing two distinct classes among the objects under inspection (2) as a function of the intensity (or luminance) of the pixels corresponding to the stored profile line: namely, a dark class (see definition above) comprising pixels of the luminance profile line below a predefined threshold, and a clear class comprising pixels of the luminance profile line above the same threshold.

[0100] In fact, pixels of clear objects (transparent objects) have high values, while pixels of dark objects have low values. Therefore, the distinction between clear and dark objects is possible through the two thresholding operations (operations) described above (the first for detecting objects acting on elevation, and the second for distinguishing between clear and dark objects acting on pixel intensity). It should be noted that dark objects with bright spots may potentially appear as clear objects in this type of processing.

[0101] The inventors have completed operational tests for prototypes of the aforementioned device and machine, and the results of these tests, illustrated as examples in FIGS. 8a through 8f, show that a nearly complete height profile of a moving object (2), in this case a bottle, can be obtained.

[0102] These drawings illustrate the process of forming a temporal accumulation of the measured profiles. The time axis is virtually displayed from top to bottom in the drawings (the image at the top of the drawing corresponds to the one taken before the one below). The last image of each sequence saved by the camera is displayed in the left part of this drawing. The interpretation of the profile lines is superimposed on the next image.

[0103] - The gray line corresponds to the profile line below the threshold to trigger object detection:

[0104] - These lines marked in white correspond to locations where the object detection threshold is exceeded.

[0105] The right portions of each of FIGS. 8a through 8f illustrate reliefs that were thus calculated and accumulated over approximately 20 consecutive time measurements. The reliefs are represented in pixels in these figures.

[0106] In the acquired final profile (Fig. 8f), the only missing surface elements are points that are simultaneously invisible as straight lines by the detection means (7) / camera and the lighting means (4) / halogen lamp + reflector assembly (e.g., protrusions (bottom of the bottle)).

[0107] If a complete profile of height (Z values ​​for all X, Y positions of the object in the image) is provided, the map can be reconstructed into the external volume of the object under inspection (2). Thus, its volume can be calculated.

[0108] This also allows the "valley" between two contacting objects (2) to be identified in the profile: and then the appropriate program can search for the local minimum height to help divide the group of adjacent objects into individual objects (2).

[0109] Of course, the present invention is not limited to the embodiments illustrated and described in the accompanying drawings. Variations are possible, particularly in terms of the formation of various elements or by substitution of technical equivalents, without necessarily departing from the scope of protection of the present invention.

Claims

Claim 1 As a device (1) for inspecting an object (2) moving in a substantially monolayer flow (F) on a flat support surface (3) moving in the longitudinal direction (D), the device (1) comprises, on one hand, at least one illumination means (4) providing an illumination beam (5) having a median plane or axis (PM) that extends across the direction of movement (D) and across all or part of the width of the support surface (3), and on the other hand, at least one detection means (7) having an optical axis or plane (AO) and an acquisition field (7'), the dimensions of which are, at best, equal to the dimensions of the illumination strip (6) in the direction across the direction of movement (D), and also, a non-illuminated area (6') of the support surface (3) located upstream and downstream of the illumination strip (6) in the direction of movement (D) and at least the illumination strip or strip At least one detection means (7) configured to encompass a volume extending over a portion (6) and optionally over the non-illuminated area (6'), and finally, means (8) for processing and evaluating signals or data provided by the at least one detection means (7) in a view for detecting the presence of the moving objects (2), and in a view for mapping their heights and / or determining their external volumes, wherein the objects are dark or black, and the intermediate plane or axis (PM) of one or more of the illumination beam(s) (5) and the optical axis or plane (AO) of the detection means (7) have a non-zero angle (AP) between them called the parallax angle, and, the image of the illumination strip (6) acquired by the at least one detection means (7) has a width equal to at least three times the resolution of the means,The device (1) is characterized in that the at least one lighting means (4) is configured and arranged such that when the lighting strip (6) created on the support surface (3) has a width greater than 10 mm in the direction of movement (D) and there is no object (2), it is defined by two opposing straight parallel edges (9, 9') that form a clear limit in the image of the lighting strip (6) and have sufficient contrast for the non-illuminated area (6') so that the lighting strip (6) can be recognized by the detection means (7) and the processing and evaluation means (8) regardless of the color of the moving object (2) to which the lighting strip (6) is applied, and one or each lighting beam (5) is composed of a beam originating from a non-coherent light source (4'). Claim 2 A device according to claim 1, characterized in that the lighting strip (6) formed on the moving support surface (3) has a width of approximately 20 mm in the direction of movement (D). Claim 3 A device according to claim 1 or 2, characterized in that the parallax angle (AP) is greater than 10°, and the illumination means (4) and the detection means (7) may or may not be located in the same part of a plane (PP) that is parallel to the illumination strip (6) and perpendicular to the support surface (3). Claim 4 An apparatus according to claim 1 or 2, wherein a single or each of the detection means (7) is a matrix camera, and the representation of the light strip (6) used in the detection means (7) and / or the processing and evaluation means (8) is composed of rows (LP), and extends from the bottom of the image provided by the detection means (7) and analyzed by the processing and evaluation means (8) when there is no object (2) affected by the light strip (6). Claim 5 An apparatus according to claim 1 or 2, wherein one or each of the lighting means (4) comprises, on the one hand, at least one non-coherent light source (4') having an elongated configuration and arranged across the direction of movement (D), and on the other hand, at least one element (4'') that focuses (concentrates) the radiation emitted by one or more of these light sources (4')(s) to create the lighting strip (6) on the support surface (3). Claim 6 A device according to claim 1 or 2, wherein one or each of the lighting means (4) provides permanent illumination to the strip (6) and comprises at least one light source (4') in the form of a straight filament extending across the direction of movement (D) and a focusing element (4'') in the form of a reflector having a partially cylindrical-elliptical section, wherein the filament (4') is positioned at one of the foci of the virtual ellipse of the reflector (4'') and the other foci are located near the support surface (3). Claim 7 An apparatus according to claim 1 or 2, wherein one or each of the lighting means (4) provides lighting scanning over the lighting strip (6) and comprises a fixed light source (4') in the form of a filament, for example, connected to a focusing element (4'') and, for example, in the form of a reflector having a partial elliptical cross section, and the lighting means (4) further comprises a movable transmission element (4''') and, for example, to create the lighting strip (6) by scanning across it by applying lighting originating from the source (i.e., light source, 4') and focused (i.e., focused) on the support surface (3). Claim 8 A machine that automatically inspects and sorts an object (2) moving in a substantially single-layer flow (F) on a conveyor, such as a conveyor belt type, which provides a moving support surface (3) for the object (2), and further comprises, wherein the machine is mounted on a conveyor and includes at least one inspection device (1) according to claim 1 or 2, and, if appropriate, includes a common illumination means (4) and at least two separate detection means (7), or a common detection means (7) and at least two separate illumination means (4), or otherwise comprises a plurality of devices (1) corresponding to at least two separate modules arranged side by side across a direction of movement (D), each comprising an illumination means (4) and a detection means (7). Claim 9 A method for inspecting an object (2) moving in a substantially monolayer flow (F) on a flat support surface (3) moving in a longitudinal direction of movement (D), implemented by a device (1), wherein the device (1) comprises, on the one hand, at least one illumination means (4) providing an illumination beam (5) having a median plane or axis (PM) that is directed toward the support surface (3) and whose intersection with the support surface (3) is composed of an illumination strip (6), and which extends across the direction of movement (D) and substantially over the entire width of the support surface (3); and on the other hand, at least one detection means (7) having an optical axis or plane (AO) and an acquisition zone or field (7'), wherein at least the illumination strip (6), a non-illuminated area (6') of the support surface (3) located on either side along the illumination strip (6), and extending above the illumination strip (6) and optionally above the non-illuminated area (6'). The apparatus comprises at least one detection means (7) configured to encompass a volume, and finally, means (8) for processing and evaluating a signal or data provided by the at least one detection means (7) in a view to detect the presence of the moving objects (2), and in a view to map their heights and / or determine their external volumes, wherein the objects are dark or black, and the apparatus is configured such that the optical axis or plane (AO) and the intermediate plane or axis (PM) of the illumination beam (5) have a non-zero angle (AP) between them called a parallax angle, and the method comprises, on the one hand, an image of an illumination strip (6) acquired by the at least one detection means (7) having a width equal to or greater than three times the resolution of the means (7), and on the other hand,A method comprising providing a lighting strip (6) on a support surface (3) that has sufficient contrast for the non-illuminated area (6') so that it can be used by the processing and evaluation means (8) and detected by the detection means (7) regardless of the color of the moving object (2) to which the lighting strip (6) is applied, and is bounded by two opposing straight parallel edges (9, 9') that are mutually distinguishable in an image, wherein one or each of the lighting beams (5) is composed of a beam originating from a non-coherent light source (4'). Claim 10 In claim 9, the method is characterized by extracting at least one profile line (LP) representing the elevation of the moving support surface (3) from an image of the lighting strip (6) by at least one appropriate data processing program executed by the processing and evaluation means (8), and comparing this profile line with a minimum threshold value for the presence of an object, and, where appropriate, using this profile line to determine the height of the object under inspection (2) point by point. Claim 11 In claim 10, the method is characterized by including analyzing the brightness of a continuous profile line (LP) of an object (2) and using this to classify the object (2) as a dark object or a transparent object. Claim 12 A method according to claim 10 or 11, wherein the method comprises accumulating consecutive profile lines of the same object (2) over time to reconstruct the apparent shape and / or compute its volume. Claim 13 A method according to claim 10 or 11, wherein the profile detection program applied by the processing and evaluation means (8) performs a search for luminance percentiles detected along each line perpendicular to the image of the lighting strip (6), and then performs an operation (operation) to threshold the profile lines. Claim 14 A method according to claim 10 or 11, wherein the processing and evaluation means (8) performs two profile detection operations (operations) to find and use two limits, namely, high and low intersections of the existing object (2) and the lighting strip (6). Claim 15 In claim 14, the method comprises determining a first profile according to an upper limit (LH) defined by a first angle (α1) and a second profile according to a lower limit (LB) defined by a second angle (α2), wherein these two limits (LH, LB) correspond to two edges (9, 9') of a light strip (6) projected onto an object (2) and are temporally offset according to the movement of the object under inspection (2). Claim 16 A method according to claim 14, characterized in that the two operations (operations) are distinguished and performed separately, and the two profiles are determined independently of each other. Claim 17 A method according to claim 14, characterized in that the method comprises executing at least one inspection device (1) according to claim 14.

Images