FR2783442A1 - Postal packet multiple packet separator method having multiple containers holding packets vertical and lower V-shaped section with lower conveyor belt section base area detecting/separating packets. - Google Patents

Abstract

The conveyor container multiple object technique has a continuous container set (G) holding packets vertically and moving along a conveyor line with a lower conveyor belt (S). Each conveyor container has a lower V-shape section. Prior to packet conveyor container loading, a threshold (1) is established. Initially, packet loading is into the V-shaped section. If more than one packet is present (T), then as the conveyor belt comes into play on the base of the packets, the excess packets are separated out and multiple packet loading is avoided.

Description

METHOD FOR DETECTING MULTIPLE PRESENCE OF OBJECTS IN

  CONVEYOR BUCKETS OF A SORTING MACHINE

  AND DEVICE FOR ITS IMPLEMENTATION

  The present invention relates to the detection of the multiple presence of flat objects in the conveying containers called "V" bottom buckets of a sorting machine. In this type of machine, the buckets are intended to transport flat objects individually between a loading location and one of a plurality of destination locations along the conveyor line, selected according to the address of the object. In general, the unitary distribution of the objects in the buckets is carried out from unstacking devices. In practice, and for various reasons including the highly heterogeneous nature of the objects treated in the main application which is postal sorting, unstacking incidents are frequent, which result in the loading of several objects instead of just one in one same bucket. And in this case, the objects unduly transported are very likely to be then misdirected, to arrive at their destination very delayed and at the end of a journey which will have brought up

  clearly the cost price of the operation.

  The present invention results from a search for a solution to this problem, and it consists of a method and a device for detecting the multiple presence of flat objects in the buckets with a "V" bottom of a sorting machine, in downstream of the loading site, the difficulty of this detection being due in particular to the fact that the objects being transported in abutment on their lower edge, the latter naturally descends into the bottom corner of the bucket, and that consequently, in the event of multiple objects in the same bucket, their lower edges will have a strong tendency to lean against each other, without

  present continuity solutions between them.

  The method according to the invention is characterized in that it comprises the following stages: (1) making the buckets cross, downstream from the loading place, a path oriented along the trajectory of the latter and constituted by the upper strand d '' an endless strip arrangement, said track running in the same direction as the buckets and at a speed equal to or slightly higher than that of the buckets, and passing inside the buckets when the latter cross it, at a slightly level higher than that of the lowest point of the interior bottom of the buckets, so that the transported objects engage on said track by their lower edges coming into contact with the upstream return means of said endless strip arrangement at a point in time position before twelve o'clock, to rise in the buckets during transit on the said track, and to be separated from one another if there are more than one in the same bucket; (2) make images of the lower edge of the objects transported during their passage on said track; and (3) process the images produced in step (2) in order to obtain an indication of the multiple presence of objects or not in the cups after each pass

audit loading station.

  To obtain significant images in step (2), the lighting is advantageously directed with a relatively grazing incidence relative to the objects transported, and from downstream to upstream relative to the direction of movement of the buckets. The image obtained according to the method and processed in step (3) thus shows the lower edge of an object in the form of a light longitudinal stripe in a dark background. A significant image of multiple presence of objects therefore comprises at least one dark intermediate furrow between two zones

longitudinal clear.

  In the application for which the invention was made, with regard to postal sorting, the heterogeneity of the objects treated is not only dimensional, but also relates to the nature of the envelopes for which the film in synthetic material is used more and more. And in the case of transparent envelopes, the production of images in step (2) requires special care to discriminate between the presence in the cups of a single ply with such a transparent envelope and the

presence of several objects.

  According to an additional characteristic of the invention, a solution is provided to this problem by providing, in addition to the lighting in normal light necessary for obtaining images, additional lighting in structured light such as laser light, intended to produce a transverse trace in the form of a line on the image of the objects. If the object transported is a fold in a plastic envelope, which may have made appear on the image resulting from normal lighting a dark furrow intermediate between two bright areas, the transverse trace resulting from said additional lighting is not interrupted at the level of said dark intermediate groove, and this indication is necessary for the erroneous indication of multiple presence of linked objects

  upon detection on the image of said intermediate groove.

  If, on the contrary, several objects are transported simultaneously in the same bucket, the transverse trace resulting from said additional lighting presents an interruption at said intermediate furrow, confirming the indication of multiple presence relative to

this last.

  According to other characteristics of the invention, said additional lighting can be emitted in the form of a flat brush, advantageously vertical, from a point below the buckets, and directed upwards and upstream by relative to the direction of movement of the buckets; and it can also be directed with a transverse obliquity with respect to the direction of movement of the buckets, in order to increase the legibility on the images of the interruptions of said trace coming from it, the interruptions then making said trace appear in the form of successive commas, of which the adjacent ends are

markedly offset.

  Other characteristics of the method according to the invention, in particular concerning image processing

  from step (3), will appear in the description

  next. The invention also relates to a device for implementing said method, comprising in combination: an endless strip arrangement immediately downstream of the bucket loading place, the upper strand of which forms, from an upstream return means, a channel oriented along the trajectory of the buckets, which track is arranged in such a way that when it is crossed by the buckets, it penetrates inside them, at a level slightly above the lowest point of the interior bottom buckets, and scrolls in the same direction as the buckets and at a speed equal to or slightly higher than that of the buckets; means for producing images of the lower edge of the objects transported during their passage on said track; and image processing means for detecting the

  multiple presence of objects in the buckets.

  These and other features and advantages of the invention will become more apparent from the

  reading of the following description, made in relation

  with the accompanying drawings, in which: FIG. 1 is a schematic side view of the mechanical and optical parts for producing images of a device according to the invention, installed in a postal sorting machine; Fig. 2 is a front view of the assembly shown in FIG. 1; Fig. 3 is a schematic side view, illustrating the operating principle of the mechanical part; Figs. 4 to 7 are views representing examples of images obtained with a device according to the invention; and Figs. 8 and 9 are graphs relating to the analysis of the images carried out in step (3) of the method

according to the invention.

  First considering Figs. 1 and 2, they therefore illustrate a device according to the invention installed in a postal sorting machine for flat objects comprising a carousel of buckets which serve a plurality of destinations downstream of a loading station. In Fig. 1, the machine is schematically represented by a block of three successive buckets G located directly downstream of a loading station not shown, and moving in the direction of the arrow F. The buckets G are intended to convey a single object at a time, which is loaded from the top and descends until it rests on the bottom by its lower edge, in the immediate vicinity of the same side for all the buckets (the right side in Fig. 2 ) in

  because of a prior teasing of the objects.

  The unloading of buckets G is ensured by retraction of the bottom which, for this purpose, consists of a pivoting flap V. For reasons of unloading efficiency, the flap V is quite inclined and therefore forms with the rear wall of the bucket a Internal corner of relatively closed angle or point P. Therefore, and as illustrated in the drawing, the objects for the most part descend into the buckets until their lower edge T comes to be placed deepest in the point P. And if inadvertently two (or more) objects have been loaded into the same bucket, in the vast majority of cases, their lower edges T will be joined against each other, without any spacing, as illustrated in the upstream bucket at

Fig. 1.

  Faced with this problem, it was nevertheless chosen, according to the invention, for the detection of multiple presence of objects in the buckets (in the sense here of several objects being together in the same bucket), to operate from below at level of the slices of the objects, by integrating in the detection process a first step consisting of an action on said lower slices, with a view to creating between them if they are several in the same bucket, a separation

  significant for the rest of the process.

  According to the invention, the separation between lower sections of objects in the same bucket is created by making them meet the upstream deflection means Ra of a horizontal channel S in an endless band arranged longitudinally under the trajectory of the buckets, directly downstream from the loading station for these, said track S, when it is crossed by the buckets, penetrating inside them, at a level slightly above the internal bottom of the buckets G at their point P, and scrolling in the same direction as the buckets G, at a speed equal to theirs otherwise

slightly higher.

  As clearly illustrated in FIG. 3 which represents the two lower slices T of the same pair of objects, first at the time of their meeting with the upstream return means Ra, then during their transit on the channel S, the engagement of the slices T on said path S will have the effect of raising them relative to the bottom of the buckets, and therefore of offering them a wider space to allow them to move away. On the other hand, by meeting the upstream return means A in a time position of about 11 hours, the adjacent sections T will be distant on the periphery thereof from a segment of rope 6 which will be kept horizontally provided that the grip on the track is satisfactory, which is the case using notched belts. The spacing 6 thus created, even if it is minimal, will be sufficient to be effectively usable in the rest of the process. It should be noted that said channel S is preferably chosen to be wide enough so that this difference is created over a significant part of the length of the lower sections of objects. Also note that said gap can be further increased by scrolling said channel S at a speed slightly

higher than that of the buckets.

  The device shown in Figs. 1 and 2 comprises the aforementioned means for spreading the lower sections of objects T, and associated means for producing images of the lower sections discarded during their passage on the track S, namely means

  lighting and a means of taking images.

  The device is mounted in a frame 1 essentially comprising two side plates 1a and 1b. It is arranged longitudinally under the trajectory of the buckets G, near the same side of the latter corresponding to a jogging alignment, and

  directly downstream of the bucket loading station.

  The endless band arrangement constituting the track S comprises two identical sets in parallel, each comprising a belt 2 passing over upstream and downstream idler rollers, respectively 3 and 4, and a lower roller 5. The horizontal upper strands of the two assemblies constituting the track S, as well as consequently the upper part of the idler rollers 3 and 4, are at a level slightly higher than that of the interior bottom of the buckets G during the crossing by them of the two assemblies, the buckets G having openings or passage notches 6 at the bottom of their rear wall and of their flap V. In the two assemblies, the drive of the belts 2 is ensured by the lower rollers 5 which, for this purpose, are mounted on the same transverse axis 7 connected to a drive means, not shown. The upper rollers 3 and 4 are in turn mounted independently of each other on the side plates 1a and 1b, so that the upward space between the two assemblies is completely clear and available for the lighting and taking pictures. As an indication, in a successful prototype, the upstream idler rollers 3 have a diameter of 120 mm, the track S being raised by about mm above the inner corner P of the buckets G. Is also visible at Fig. 1, a sliding ramp 8 leading from upstream to downstream going up on the periphery of each upstream roller 3, which has the function of contributing to the passage of objects on the rollers 3, and which is especially useful for fine and light objects

to avoid kickbacks.

  The lighting means comprise a light source 10, visible in FIG. 1 but not shown in FIG. 2 for clarity of the drawing. Source 10 is located at the top and near the downstream end

  of the device, centrally between the two sets.

  It is directed longitudinally upwards and upstream, at an angle of approximately 20 with the horizontal, on the front part of the track S. To this lighted area corresponds an aperture 11 in the bottom of the buckets G, consisting of indentations in their rear wall and their flap V. The image taking means 15, advantageously being a digital camera or a laser sensor, is placed vertically, just below the illuminated area , the development being made on the plane of the track S. It is triggered by a detector 16 placed slightly upstream of the lit area to signal the arrival of the buckets G. The detector 16 is advantageously a contactor, actuated at pssage buckets by a stud

attached to the side of these.

  This main lighting is associated with structured lighting, for example in the form

  of laser line, coming from a source 12 placed above

  below the main source 10. This light is directed on the same zone as the main lighting, at an angle of about 40 with the horizontal, but instead of being longitudinal, it has a certain transverse obliquity, fixed in the example shown in 25. The skate is advantageously a

  vertical flat brush, preferably divergent.

  Figs. 4 to 7 represent the images which one obtains with a device such as that which comes

to be described.

  The image in Fig. 4 is representative of an empty bucket. The two clear upper and lower lines 20 identified represent the two sides of the window 11 in the bottom of the bucket, and indicate that the lighting has worked well. The absence of these lines indicates a defect

  of operation such as absence of lighting.

  The image in Fig. 5 corresponds to the presence of a single object in the relevant bucket, appearing under

  the shape of a single clear transverse band.

  The image in Fig. 6 corresponds to the presence of two objects appearing in the form of two clear transverse bands 11, clearly separated by a dark intermediate groove or border representative of the interval between lower sections of the two objects. Finally, the image of FIG. 7 corresponds to the particular case of a single object (transparent envelope) whose image resulting from the main lighting source does not differ sufficiently from an image of multiple presence to allow discrimination. On the other hand, it can be seen that this discrimination becomes evident in view of the oblique transverse trace 21 resulting from the additional lighting, because it is not here interrupted in the crossing of the dark intermediate groove normally indicative of an interval between objects, and is very clearly differentiated from trace 21 of FIG. 6 similarly resulting from the additional lighting and which itself consists of two segments spaced from each other at the level of the intermediate groove, and the adjacent ends of which are clearly offset in the

transverse direction.

  The processing of these images will now be described, the purpose of which is to control the buckets after loading, with a view to conveying, to destinations of the sorting machine reserved for this purpose, the abnormally transported objects in the same bucket. It is performed on digital images in which each of the pixels has a level value of

gray or luminance between O and 255.

  Step 1: It consists of determining whether the image to be processed contains object presence information. We can deduce that there is presence of object if the image presents

  locally a higher luminance area.

  In order to detect such a higher luminance area, the algorithm performs a horizontal profile, Fig. 6, and determines whether it has a peak, as illustrated in the graph of FIG. 8 o are plotted on the abscissa the horizontal coordinates of the pixels and in

  ordered the luminance or gray level values.

  As shown in Fig. 6, detection begins at mid-height of the image and takes place downwards. If on arrival at the bottom of the image, there was no detection of a peak, this indicates the absence of an object in the bucket, or else a lighting defect, and the distinction may be made between these two eventualities by providing in the algorithm a high trace detection and

  low 20 appearing in the image of FIG. 4.

  Step 2: It is a question of looking for minima in a peak

detected in step 1.

  The previous profile, Fig. 8, made it possible to determine whether or not there is an object or objects in the bucket corresponding to the image processed. This profile, with a view to optimizing the processing time, is advantageously produced by summing a few horizontal lines, and it can therefore have sudden variations due to the noise of the system. In order to eliminate them, a new profile corresponding to a greater number of horizontal lines, for example fifty, is produced, at the place where the presence of at least one object was detected in step 1.

  sufficiently smooth for further treatment.

  From the profile obtained, all the local minima capable of representing the separation between two objects are determined. To be accepted, the minima must meet three criteria: (a) the dynamics of the ridges surrounding the minima must be at least 30; the dynamics being the difference between the luminance values of the peak and the minimum; (b) the slopes on either side of the minimum to the ridges which surround it must be sufficiently large; (c) the measured values of the two share slopes

  and other parts of the minimum must be of the same order.

  The diagram in FIG. 9 has a profile with a peak where the first two minima A and B meet the three criteria (a), (b) and (c) and will therefore be used for further processing, while minima C

will be ignored.

  Step 3: It consists of a contour tracking on the original image. From the coordinates of the minima retained on the profile and the position of the profile on the image, it is possible to find where the minima is in the original image. In this step, it will be checked if the minimum adopted which indicates a difference

  potential between two objects is sufficiently large.

  To this end, the contour of the minimum space is followed and a measurement of its length, of

the way illustrated in FIG. 6.

  In a first phase of this step, it is sought line by line, from the place where the profile was made, the presence of a minimum. The search area is restricted according to the coordinates of the minima and the two peaks surrounding it (maxima) found on the starting profile. On each line, we look for the minimum and the two maxima, then we measure the difference between the two maxima, which must be sufficiently significant (for example greater than 5 pixels) to

to be taken into account.

  The second phase consists in determining the point of the lowest luminescence minimum, which will be used to determine the starting point for processing the next line, exactly above or below depending on the direction of the operation. This point immediately above or below must also have a very low luminance. Otherwise, a penalty is assigned to the area explored. The point of lower luminance is also used for the detection of a trace 21 resulting from additional lighting in structured light (such as laser light), which detection will be carried out in step

following treatment.

  In the third phase, a score is calculated, which determines whether the area explored is sufficiently significant of a space between objects. To do this, the number of weakest luminescence points for which the above condition is accumulated

  of difference between maxima beyond the minimum is satisfied.

  This score is decreased by the number of penalties, and it is

  normalized with respect to the length of the contour.

  It is concluded that there are multiple objects on the double condition that the score is greater than 70% and

  that the number of penalties does not exceed 20.

  Step 4: This involves analyzing the aforementioned trace 21 resulting from the additional lighting to confirm or invalidate the indications of multiple presence of objects resulting from the previous step. If in the second phase of the previous step, a first point of exploration of a line is detected which should be a minimum with a luminance value significantly higher than that of the point immediately above or below (for example greater than half the sum of the values of this adjacent minima and one of its maxima),

  the algorithm searches for a trace 21.

  For this, an exploration takes place to the right and to the left of said first point detected to search for a saturated zone of small width. If the criterion is verified to the left and to the right of said first point, the algorithm stops the exploration of the contour considered, concluding that there is no interruption of the trace 21 at the level of this contour, and that consequently, the latter does not correspond to a space between two objects. The coordinates of the ends of trace 21 are

  saved for the next step.

  Step 5: This is the final analysis. The image has been subdivided into two independent upper and lower parts which have each undergone the processing of steps 1 to 4. And the synthesis of the results obtained on the two parts of the image is now carried out. The indication of multiple presence of objects is finally declared if it is cumulatively satisfied with three conditions: - that there has been detection of multiple presence in the two parts of the image; - that the grooves or boundaries between objects detected on the two half-images are geographically close (for example, less than twenty pixels apart); - that the detected path is not between

  the two ends of a trace 21 detected.

  Note that trace 21 is only found in one of the parts of the images, and that its search is therefore

performed only in this part.

Claims (12)

  1) Method for detecting the multiple presence of objects in the cups (G) with a "V" bottom of a sorting machine for flat objects, in which the objects are transported upright on their lower edge in the cups (G ), which method is characterized in that it comprises the following steps: (1) making the buckets (G), downstream of the loading place, cross a track (S) oriented along the trajectory of the latter and constituted by the upper strand of an endless band arrangement, the track (S) running in the same direction as the buckets and at a speed equal to or slightly higher than that of the buckets, and passing inside the buckets when these cross it, at a level slightly higher than that of the lowest point (P) of the interior bottom of the buckets, so that the transported objects engage on the track (S) by their lower edges (T) coming into contact with the upstream return means (Ra) of said arrangement in b ande endless at a point in the time position before twelve o'clock, to rise in the buckets (G) during transit on the track (S), and be separated from each other if they are several in the same bucket; (2) make images of the lower edge (T) of the objects transported during their passage on the track (S); and (3) process the images produced in step (2) in order to obtain an indication of the multiple presence of objects or not in the cups (G) after each passage at the place of loading.   2) Method according to claim 1, characterized in that to carry out the images in step (2), there is provided a directed lighting with a relatively grazing incidence relative to the objects transported, and from downstream to upstream with regard to   the direction of movement of the buckets.   3) Device according to claim 1 or 2, characterized in that to produce the images in step (2), there is provided in addition to lighting in normal light necessary for obtaining images, additional lighting in structured light, intended to produce a transverse trace (21) in the form of a line on the image of objects.   4) Method according to claim 3, characterized in that said additional lighting is a source of laser light.   ) Method according to claim 3 or 4, characterized in that said additional lighting is emitted in the form of a flat brush, from a point below the buckets (G), directed upwards and upstream with respect to the direction of movement of the buckets, and with a transverse obliquity with respect to the direction of movement of the buckets, in order to increase the legibility on the images of the interruptions of said trace (21) coming from it, the interruptions then revealing said trace in the form of successive commas whose ends   adjacent are clearly offset transversely.   6) Method according to one of claims 1 to 5,   characterized in that the image processing step (3) itself comprises the following steps: - detecting information on the presence of an object in the bucket corresponding to the image considered, this information being in the form of a luminance peak on a horizontal profile produced on the image; - in the event of positive detection, search for minima in a detected peak and retain those which are potentially representative of a separation between two objects; - perform a contour tracking on the image considered from the coordinates of a selected minimum, and determine if this contour is potentially indicative of a multiple presence of objects; - analyze all the results to declare   whether or not there is multiple presence of objects.   7) Method according to the preceding claim, itself dependent on claim 4, characterized in that the step (3) of image processing itself comprises an additional step relating to the analysis   of the trace resulting from said additional lighting.   8) Method according to claim 6 or 7, characterized in that in step (3) of image processing, the images considered are subdivided into two parts each subjected to the processing preceding the final analysis, the comparison of the results concerning the two parts of the image being taken into counts in the final analysis.   9) Device for implementing the method   according to one of the preceding claims, characterized   in that it comprises in combination: an endless strip arrangement immediately downstream from the place of loading of the buckets (G), the upper strand of which forms, from an upstream return means (Ra or 3), an oriented path (S) along the trajectory of the buckets (G), which track (S) is arranged in such a way that when it is crossed by the buckets (G), it penetrates inside them, at a slightly level above the lowest point (P) of the inner bottom of the buckets, and scrolls in the same direction as the buckets (G) and at a speed equal to or slightly higher than that of the buckets; means for producing images of the lower edge (T) of the objects transported during their passage on said track (S); and image processing means for detecting the   multiple presence of objects in the cups (G).   ) Device according to claim 9, characterized in that said endless strip arrangement comprises between the side walls (la, lb) of a frame (1), two identical sets in parallel, each comprising a belt (2) passing over upstream and downstream return rollers (3, 4), the horizontal upper strands in the two assemblies forming track (S).   11) Device according to claim 10, characterized in that each set of said endless strip arrangement comprises a lower roller (5) which is motor, the two rollers (5) being mounted on the same transverse axis (7) connected to a drive means, while the upstream and downstream idler rollers (3, 4) are mounted independently of each other   others on the walls (la, lb) of the frame (1).   12) Device according to claim 10 or 11,   characterized in that the belts (2) are notched.   13) Device according to one of claims 10 to   12, characterized in that it comprises a ramp (8) for accessing objects on each upstream return roller (3).   14) Device according to one of claims 10 to   13, characterized in that it comprises a light source (10) disposed at the top and near the downstream end of the frame (1), and a structured light source (12) placed below the main source (10), both directed from bottom to top and   downstream upstream, towards the same area of the track (S).   ) Device according to claim 14, characterized in that the source (12) emits laser light, in the form of a substantially flat brush   vertical and oblique transversely.