FR2738799A1 - MACHINE AND METHOD FOR SORTING AND COMBINING INDIVIDUAL ARTICLES FOR PREDETERMINED QUANTITY PACKAGING - Google Patents

Abstract

The system uses an artificial vision unit (3) including a matrix camera (7) whose zone of vision (9) encompasses at least a part of the collection means (1). The vision system (3) permits the referencing and counting of the articles passing through the vision zone (9). A retention comb (4) temporarily restrains the articles driven by the collection system (1). The comb (4) is placed on the downstream side of the vision zone (9), and has a number of teeth (10, 11) each of which can take up two positions, that is to say a rest position and a retention position, in which the teeth hinder the movement of the articles or let them pass freely. The vision system (3) sends control data to the comb (4) so that the comb's teeth are lowered from the rest position to the retention position, until the end of the inspection cycle. Lowering of certain teeth is delayed so as to selectively authorise the passage of the last articles (32) which enables the matching of the number of articles (N) with the predetermined quantity (NO) for the current inspection cycle.

Description

Machine and method for sorting and grouping individual articles

  for their conditioning by predetermined quantity.

  The field of the invention is that of packaging articles derived for example from an industrialized production line. More specifically, the invention relates to a machine and a method for sorting and grouping individual articles,

  way that they can be conditioned by predetermined quantity.

  There are many applications in which articles are grouped in batches to be packaged together, for example in sachets. Respect for a

  given number of items in each lot will not hurt the customer or the manufacturer.

  The invention applies in particular, but not exclusively, to the sorting and grouping of frozen products of pastry, bakery or pastry (croissants for example) in view of their bagging, each bag to contain a given number of frozen products. In this case, the sorting and grouping operation and the bagging operation are carried out in a conditioning room, located between a deep-freezer and a cold storage chamber. Frozen products arrive

  usually in the packing room arranged loose on a conveyor.

  Usually, the freezer and the cold storage room are at a very low temperature (usually - 40 C), while the conditioning room is at a higher temperature (usually + 10 C). Indeed, it must be possible, inside this room conditioning, to work people and machines with electronics. For conservation reasons, frozen products can only be kept in the packaging room for a few minutes and therefore sorting, grouping and bagging operations must be carried out very quickly. The conditioning operation is carried out by a suitable machine, for example a bagger in the aforementioned application. Since the invention relates solely to the sorting and grouping operation, the packaging operation is not discussed.

  in more detail later in this description.

  In the state of the art, several techniques are known that make it possible to

  to perform the sorting and grouping operation.

  A first known solution is to use a centrifuge of the type comprising a slightly conical tray honeycombed around its periphery. The tray is fed vertically, from above, and rotates slowly. Its low slope combined with its rotation ensures a flow and distribution of articles to the cells. A conveyor for evacuation of articles to a packaging device is disposed under the centrifuge tray. Each cell, when it passes over the conveyor belt, opens to drop the item it contains. Thus, the articles are deposited one by one on the evacuation belt and it is therefore easy to count them. When the quantity

  predetermined set of items to put in a bag is reached, the centrifuge is blocked.

  As soon as the bagging in progress is finished, the following conditioning cycle can begin: the centrifuge is unlocked, which releases the items one by one on the conveyor belt. This first known solution makes it possible to always have the right number of articles in each sachet. On the other hand, it does not allow, because of its principle of release of articles one by one, to reach high rates. In addition, the centrifuge can only be used with articles of a given size, which is imposed by the

dimension of the cells.

  According to a second known sorting and grouping technique, the articles arrive on a single conveyor belt. At a junction point, two doors make it possible to direct the articles towards one or the other of two parallel packaging installations. A vision system, including in particular image processing means and a linear camera placed upstream of the branch point between the two installations, makes it possible to detect and count incoming articles. All items that pass are first directed to one of the packaging facilities. When the number of past articles reaches the predetermined amount of articles to be packaged together, the delivery conveyor belt is stopped after the last items to fill the desired number have passed the branch point. Then the doors are moved so that all subsequent items are directed to the other packaging facility, and the supply conveyor belt is restarted. Thus, the two conditioning plants are supposed to receive

  alternatively a predetermined quantity of articles.

  This second known technique of sorting and grouping has the major disadvantage of not being able to process a continuous stream of articles whereas almost all the production flows at the output of the industrialized production lines are precisely of the continuous type. Indeed, the periodic stop of the supply treadmill in the case of a continuous stream of articles would result in an unacceptable accumulation of items on the conveyor belt. Due to the use of a linear camera, counting of items is fraught with errors. Indeed, if the articles move on the carpet when the latter advances, on the reconstituted image from a succession of lines of pixels, these articles appear

  either larger or smaller than they actually are.

  Yet another disadvantage of this second known technique of sorting and grouping is not to ensure an exact number of articles per conditioning cycle. Indeed, due to mechanical constraints, it is difficult to perfectly synchronize the movement of the doors with the passage of the last article

  to reach the predetermined quantity.

  Moreover, if several articles are contiguous on the conveyor belt, and if the predetermined quantity is reached with only a part of them, the doors can not separate them and all of them lead to the installation associated with the conditioning cycle. course. In the latter case, the installation associated with the current conditioning cycle receives a number of items too high, and the other installation, associated with the next conditioning cycle, receives a number of items too low. To avoid this, a person must read on a display screen the actual number of items received by each packaging facility, and remove or manually add items to reach the predetermined amount. Such a manual operation of adjusting the number

  of articles greatly increases the cost of packaging.

  The purpose of the invention is notably to overcome these various disadvantages of

the state of the art.

  More specifically, it is an object of the present invention to provide a machine and method for sorting and grouping individual articles to ensure that the number of articles packaged together is always equal to or greater than

  one unit maximum to a predetermined quantity.

  The invention also aims to provide such a machine and such a method that are configurable, so as to be used with any type of article. Another object of the invention is to provide such a machine and such a method.

  that are self-adapting according to the nature of the articles.

  Yet another object of the invention is to provide such a machine and such a method with a very flexible operation and to achieve rates

very high.

  A complementary object of the invention is to provide such a machine and a

  such a method to reduce packaging costs.

  These objectives, as well as others which will appear later, are achieved according to the invention using a machine for sorting and grouping individual articles in view of their conditioning by predetermined quantity during conditioning cycles. successive, said machine being of the type comprising in particular means for feeding said articles to a conditioning device, said machine comprising: an artificial vision system comprising in particular a camera whose zone of vision includes at least a portion of said means of vision; brought, said artificial vision system for identifying and counting articles passing in said viewing area; a comb of temporary retention of articles driven by said feeding means, said comb being placed downstream of said viewing zone and comprising a plurality of teeth each of which can take two positions, namely a rest position and a position of retention, wherein said tooth does not impede and impede respectively the passage of articles; said artificial vision system delivering comb control information so that the comb teeth are lowered from the rest position to the retention position until the end of the current conditioning cycle, the lowering of certain teeth being delayed so as to selectively allow the passage of the last item or articles to supplement the number of items to said predetermined quantity for the

conditioning cycle in progress.

  Thus, the general principle of the invention is to enslave a temporary retention comb articles to a control information generated by an artificial vision system. For each conditioning cycle, the comb teeth are selectively lowered to pass the correct number of articles. The comb holds the other items, so that the conditioning cycle ends with the right number of items, and

  until the conditioning device is ready to start a new cycle.

  Unlike the mechanical arm of the second known solution of the state of the art, the comb allows to pass some items and block the others, so as to achieve, for the current conditioning cycle, the right number of items (o to approach as much as possible by excess). This is possible thanks to the teeth that can be lowered selectively and to take into account the disposition of articles

on the means of supply.

  The machine of the invention can operate at very high speeds since no interruption is necessary. Indeed, the supply means transporting articles permanently to the packaging device, and the comb ensures a merely temporary retention of some of these items, at the end of the cycle.

  conditioning, so that the right number of articles can be passed.

  The machine of the invention does not require the presence of a person to manually adjust the number of articles for each conditioning cycle. By

  therefore, packaging costs are reduced.

  If the pitch of the comb, that is to say the spacing between two successive teeth, is sufficiently small, the machine of the invention can operate regardless of the type and size of articles. Indeed, only the artificial vision system must be configured according to the nature of the items to be identified and counted. In other words, the system of

  Artificial vision makes the machine of the invention very versatile to use.

  Advantageously, said machine comprises a temporary retention barrier of articles driven by said supply means, said barrier being placed downstream of said temporary retention comb, lowering said barrier until the end of the conditioning cycle in progress being delayed with respect to the lowering of the comb, so as to allow the passage of said one or more last articles making it possible to top-up, the lowering of said barrier allowing said comb to be raised before the end of the current conditioning cycle. Thus, according to this variant, there are two locking times of the articles: - firstly, the comb ensures a first selective blocking of articles, allowing the last item (s) to reach the right number of articles for the conditioning cycle in progress; - then the barrier ensures a second non-selective blocking of the released articles

  by the comb as well as all those arriving thereafter.

  In this way, the comb can be reassembled before the end of the current conditioning cycle, which avoids an accumulation of articles in the viewing area, and therefore errors in identifying and counting articles. Indeed, the viewing zone is generally close enough to the comb, upstream, to reduce the risk of movement of articles between the viewing zone and the comb, such movements being

  cause errors in the articles actually retained by the comb.

  Preferably, said machine comprises, upstream of said zone of vision, a brush placed above said supply means and rotating in the direction opposite to the driving direction of said supply means, so as to eliminate any superimposition

  articles on said supply means.

  Preferably, said feed means comprise, upstream of said zone of vision, an acceleration zone, for spreading said articles on said

supply means.

  The brush and the acceleration zone, which are two means that can be combined, facilitate on the one hand the tracking and counting of the articles by the vision system.

  artificial, and on the other hand the selective blocking of certain articles by the comb.

  Advantageously, said camera is a matrix camera. In this way, the articles are not distorted on the images, even if they move on the means

feed.

  The invention also relates to a method of sorting and grouping individual articles in view of their conditioning by predetermined quantity during successive conditioning cycles, said articles being driven to a packaging device by supply means, said method comprising , for each conditioning cycle, the following steps: - identification and counting, by an artificial vision system comprising in particular a camera whose area of vision includes at least a portion of said supply means, articles passing through said zone of vision ; and, when said predetermined quantity of articles is reached for said conditioning cycle: calculating, by said artificial vision system, control information of a temporary retention comb of articles driven by said supplying means said comb being placed downstream of said viewing zone and having a plurality of teeth each of which can take two positions, namely a rest position and a retention position, in which said tooth does not obstruct and impede respectively the passage of the articles, said comb control information being such that the lowering of certain teeth is delayed so as to selectively allow the passage of the last item or articles for adding the number of articles to said predetermined quantity; - lowering the comb teeth, from the rest position to the retention position, until the end of the current conditioning cycle, according to said

comb control information.

  Advantageously, said steps implemented in said artificial vision system for each conditioning cycle constitute an iterative process, each iteration is linked to the previous one and comprises: said step of locating and counting the articles passing in the vision zone and, only when said predetermined quantity of articles is reached for the current conditioning cycle:

  said step of calculating comb control information.

  Thus, the artificial vision system operates according to an iterative process whose iterations are not repeated at a predetermined period but in a continuous manner, each new iteration being linked to the previous one. Moreover, not all iterations have the same duration. Indeed, some include only the step of locating and counting items, and are therefore shorter than others which include more

  the step of calculating a control information of the comb.

  Advantageously, said step of locating and counting articles passing through the viewing zone comprises the following steps: storing a digital image taken by said camera; calculating the number of new articles that have appeared in the viewing zone, by comparing the last image stored with the previous stored image; calculating the cumulative number of articles passed in the viewing area since the beginning of the current conditioning cycle; comparing said cumulative number of articles with said quantity

predetermined number of articles.

  Thus, items passing in the viewing area are counted when they first appear on an image. If articles have not had time to traverse the entire viewing area before taking the next digital image, they are still present in the next image, but at different locations. The artificial vision system

  recognizes them and does not take them into account again.

  It should be noted that a binarization step of the stored digital image can

  advantageously preceding the calculation steps relating to the image.

  Preferably, said step of calculating a comb control information comprises the following steps: searching for groups of articles on said last stored image, each group comprising one or more adjacent articles and constituting an indivisible assembly by said comb; calculating, according to a predetermined strategy, a selection of at least one group and an associated lowering order of the teeth selectively permitting the passage of said group of the selection, the item or articles of said group of the selection constituting said or said last items to add the number of items to the predetermined amount. Thus, all the articles present on an image are distributed in one or more groups. Each group of articles appears for the comb as a single element that it can only pass or retain as a whole. The concept of group of articles is directly related to mechanical constraints, such as in particular the spacing and the inertia

  teeth of the comb, and the driving speed of the feed means.

  During the step of calculating a selection of one or more groups and a lowering order of the teeth, certain mechanical stresses (in particular those mentioned above) are taken into account as well as an authorized clearance of movement of the teeth. articles on ways

  of feeding, between the zone of vision and the comb.

  The predetermined calculation strategy consists, for example, in calculating all the possible possibilities with the groups found in the previous step, and in choosing the one that leads to an equal or nearest number of articles by excess of the predetermined quantity. Preferably, said step of calculating the number of new articles that have appeared in the viewing zone comprises the following steps: searching for tasks on said last stored image, each task comprising one or more adjacent articles and being seen by said vision system artificial as having a unique outline; determination of the new tasks that have appeared in the viewing zone, by comparing the tasks of the last stored image with those of the previous stored image; calculating the number of new articles contained in said new tasks, the number of new articles contained in each new task being obtained by dividing the surface of said task by a surface

predetermined article.

  Thus, all the articles present on an image are distributed in one or more tasks. A task may only include one article, if it is isolated from other articles. All items that are attached to each other belong to the same task, because the same contour allows to encompass all.

  Advantageously, each group consists of an integer number of tasks.

  In this way, the number of articles in a given group is easily deduced from

  number of articles in each of the tasks in that group.

  Preferably, said method comprises, following said step of lowering the teeth of the comb and when said predetermined quantity of articles is reached for the current conditioning cycle, an additional step of lowering a temporary retention barrier of articles driven by the supply means, said barrier being placed downstream of said temporary retention comb, the lowering of said barrier until the end of the current conditioning cycle being delayed with respect to the lowering of the comb, so as to authorize the passage of said one or more last articles making it possible to top-up, the lowering of said barrier allowing the raising of said

  comb before the end of the current conditioning cycle.

  Other features and advantages of the invention will appear on reading the

  following description of a preferred embodiment of the invention, given as a

  indicative and non-limiting example, and the accompanying drawings, in which: - Figure 1 shows a schematic top view of a particular embodiment of the sorting and grouping machine of the invention; - Figure 2 shows a partial perspective view of the machine of Figure 1; - Figures 3A to 3E show several successive views corresponding to intermediate states of the machine of Figures 1 and 2, in an example of operation; FIG. 4 presents a block diagram of one embodiment of the sorting and grouping method of the invention; FIGS. 5, 6 and 7 each show in detail one of the steps of the method of FIG. 4; and

  - Figure 8 illustrates the concepts of group and task.

  The invention therefore relates to a machine and a method for sorting and grouping individual articles in view of their conditioning by predetermined quantity during

  successive conditioning cycles.

  In the embodiment shown in FIGS. 1 and 2, the machine of the invention comprises: means 1 for feeding articles 2 to a packaging device (not shown), in the direction of the arrow referenced 6 ; - an artificial vision system 3; a comb 4 for temporary retention of articles;

  a barrier 5 for temporary retention of articles.

  Articles 2 are, for example, frozen pastries and the

  packaging is for example a bagger.

  A treadmill is for example the means I of supply.

  The artificial vision system 3 comprises in particular a camera 7 and

  means 8 of image processing. The camera 7 is for example a matrix camera.

  Thus, even if items are not stationary on the treadmill, they do not appear

not distorted on the images.

  The camera 7 is fixed and located opposite the treadmill 1. The viewing zone 9 of the camera 7 includes a portion of the treadmill 1 so that all the articles 2 pass in this viewing zone 9. Each image taken by the camera 7 has for example 000 pixels. The image processing means 8 are for example made under the

  form of a wired card, so as to increase the speed of treatment.

  As explained in detail in the following description, the vision system

  artificial 3 allows to identify and count the items passing in the viewing area 9, that is to say all the items driven by the treadmill 1 to the device of

conditioning (not shown).

  Furthermore, the artificial vision system 3 delivers information 14 of

  control of comb 4 and control information of barrier 5.

  The comb 4 is placed downstream of the viewing zone 9 of the camera 7, that is to say between the viewing zone 9 and the conditioning device (not shown). It comprises a plurality of teeth 10, 11. The teeth 10, 11 are for example cylinders. The pitch of the comb 4 is for example equal to half the width of the smallest articles. Each tooth 10, 11 may each take two positions, namely: a rest position (or high position), in which it do not obstruct the passage of articles; - a retention position (or low position), in which it impedes the

passage of articles.

  In FIG. 1, the teeth 11 in the rest position are represented by a circle,

  while the teeth 10 in the retention position are represented by a black disc.

  The barrier 5 is placed downstream of the comb 4. It can take two positions, namely a rest position and a retention position. Unlike comb 4, whose teeth 10, 11 are independent and can not be lowered simultaneously, the

barrier 5 is monobloc.

  Optionally, and as shown in FIG. 1, the machine of the invention may comprise a brush 12 and / or the conveyor belt 1 may comprise a zone

acceleration 13.

  The brush 12 is placed above the conveyor belt 1, upstream of the viewing zone 9. It rotates in the opposite direction to the driving direction of the conveyor belt. It makes it possible to suppress the possible superpositions of articles, and thus to make all the articles

  visible by the artificial vision system 3.

  The acceleration zone 13 of the treadmill 1 is situated upstream of the viewing zone 9 of the camera 7. It makes it possible to spread the articles on the treadmill 1, and thus to facilitate their discernment by the artificial vision system. 3. For example, the treadmill 1 can pass, in the acceleration zone 13, a speed of 10m / min to m / min. The operation of the machine of the invention shown in FIGS. 1 and 2 is now explained in relation with FIG. 4, which presents a block diagram of a

  embodiment of the method of the invention.

  This method comprises, for each conditioning cycle, the following steps: - marking and counting (41), by the artificial vision system, articles passing in the viewing zone 9; comparing (42) the number N of items already passed for the current conditioning cycle with the predetermined quantity of items to be reached; and when this predetermined quantity of articles is reached: calculation (43), by the artificial vision system 3, of the control information 14 of the comb 4; lowering (45) of the barrier 5 as a function of the control information of the barrier 5, until the end of the current conditioning cycle, and raising of the comb 4;

- raising (46) of the barrier 5.

  The steps 41 and 42 implemented in the artificial vision system 3 constitute an iterative process comprising one (41) or two (41 and 42) steps, depending on whether or not the predetermined quantity of articles is reached. Each iteration of the process implemented in the artificial vision system 3 has for example a duration of approximately 90 ms, if it comprises a single step 41, or 150 ms, if it

has two steps 41 and 42.

  The control information 14 of the comb 4 is such that the lowering of certain teeth is delayed, so as to selectively allow the passage of the last item or articles for adding the number N of articles to the predetermined quantity No The control information of the barrier 5 is such that the lowering of the barrier 5 is delayed with respect to the complete lowering of the comb 4, so as to allow the passage of the last article or articles making it possible. extra (and that the comb 4 also let pass). The instant of lowering of the barrier 5 is for example calculated from the moment of complete lowering of the comb 4, taking into account the speed of the treadmill 1 and the distance between the comb 4 and the barrier 5, so that the last items that the comb 4 has passed pass the barrier 5 before the latter is lowered. It should be noted that the comb 4 can be reassembled before the barrier 5 is lowered. Indeed, the minimum duration during which the comb 4 must remain in the lowered position is that which allows to sufficiently distance the last item of a cycle and the first item of the next cycle, so that the barrier 5 has time to go down between the

  moments of passage of these last and first articles of two successive cycles.

  An example of operation is now presented in relation to FIGS. 3A to 3E, which show different views from above of the machine, corresponding to

  successive operating instants.

  In each figure (3A to 3E), there is shown the treadmill 1, the viewing zone 9, the comb 4, the barrier 5, and the individual articles 2. Each zone of the comb 4 is represented either by a circle, if it is in the rest position, either by a black disk, if it is in the retention position. Barrier 5 is shown in dotted line, if it is

  in the rest position, either in full line, if it is in the retention position.

  In order to be able to distinguish them easily, articles 2 were marked with a cross

  affected by the current conditioning cycle.

  FIG. 3A situates itself in the case where the artificial vision system, having found that the number N of articles passed in the viewing zone was equal to the predetermined quantity No, calculated that this predetermined quantity No was reached by

  example by letting the three articles referenced 32 and blocking the other articles.

  For this purpose, some teeth 31 of the comb 4 have not been lowered at the same time as

the others 33.

  After the articles 32 for filling are passed, all the teeth of the comb 4 are lowered (see Fig. 3B) and ensure the retention of all items not concerned by the current conditioning cycle (that is, to say, in Figure 3B, those

not marked with a cross).

  Then (see Fig. 3C), in order to avoid an accumulation of articles in front of the comb 4 which could reach the near vision zone 9, the barrier 5 is lowered, and the teeth of the comb 4 are raised. This barrier 5 (see Fig. 3D) allows to retain all articles

  not concerned by the current conditioning cycle.

  Finally (see Fig. 3E), when the conditioning cycle in progress is finished, the

barrier 5 is raised.

  It should be noted that, in a simpler embodiment of the invention, the machine has no barrier 5. In this case, the comb 4 is lowered until the end.

  of the conditioning cycle in progress.

  FIG. 5 shows the step 41 for identifying and counting the process of FIG. 4. This step 41 comprises, for example, the following steps: storing (51) a digital image taken by the camera 7 and representing the viewing zone 9 at a given moment. Take for example an image every 15 cm of progress if the viewing zone 9 is about 80 cm long; binarization (52) of the stored image; calculating (53) the number of new items that appeared in the viewing area 9; calculating (54) the cumulative number N of articles passed in the viewing zone since the beginning of the current conditioning cycle, by adding the

  previous cumulative number with the number of new items.

  As shown in FIG. 7, the step 53 of calculating the number of new items that appeared in the viewing zone 9 comprises, for example, the following steps: searching (71) tasks on the last stored image; determining (72) the new tasks that have appeared in the viewing area 9 by comparing the tasks of the last stored picture with those of the previous stored picture; - calculating (73) the number of new articles contained in the new tasks. By task is meant here a set of one or more adjacent or contiguous articles, which is seen by the artificial vision system 3 as having a single contour. In other words, the artificial vision system does not know how to distinguish each article individually from a task and therefore treats it as an element as such, having coordinates (for example minimal and

  maximum, and minimum and maximum ordinates) and a surface.

  To perform the comparison of the tasks, from one image to another, the artificial vision system 3 can for example manage a list of tasks, associating with each task its surface, its coordinates and the number of products it contains. . To find a task from one image to another, we consider that the task undergoes essentially a translation movement. However, one can admit a movement capacity of a

  other type, as well as a slight surface variation.

  Calculating (73) the number of new items contained in each new job is to divide the area of the new job by a predetermined article area. In other words, an article is defined as an element having a given surface. One can also perform calculations with article surfaces

  minimum and maximum, so as to approximate the reality better.

  FIG. 6 shows in detail the step 43 for calculating the control information of the comb of FIG. 4. This step 43 comprises, for example, the following steps: - searching (61) groups of articles on the last stored image; calculation (62), according to a predetermined strategy, of a selection of one or more groups and of a lowering order of the teeth such that it authorizes the passage of the selected group or groups. The article or articles of the selected group (s) constitute the last item (s) for adding the number N of articles to the predetermined quantity. No. By group means here a set of one or more adjacent or contiguous articles. , which is indivisible by the comb 4. In other words, the comb 4, due in particular to the inertia of the teeth, the speed of the treadmill or even the authorized game of movement of the articles, does not allow to leave to pass a part of the articles of the group and to retain the others: he can treat all these articles only as a same

  together, that he holds back or lets pass.

  In order to simplify the calculations of the number of articles per group, we consider

  example that a group consists of an integer number of tasks.

  Thus, as illustrated in FIG. 7, the artificial vision system sees tasks 81, which it groups together in groups 82 as a function, in particular, of the mechanical stresses (inertia of the teeth, not of the comb, speed of the treadmill), of the authorized clearance movement of tasks, and the layout and size of tasks. The artificial vision system 3 does not distinguish the articles within the tasks, these articles being simply defined,

  in a general way, by a surface and possibly a capacity of movement.

  To calculate a selection of one or more groups and a corresponding lowering order of the teeth, a strategy may consist in calculating all the possibilities, and in retaining that which leads to a number N of articles equal to the predetermined quantity No, or

  well to a number N that approaches it the most by excess.

Claims (10)

  1. Machine for sorting and grouping individual articles in view of their conditioning by predetermined quantity during successive conditioning cycles, said machine being of the type comprising in particular means (1) for feeding said articles (2) to a device conditioning device, characterized in that it comprises: - an artificial vision system (3) including in particular a camera (7) whose viewing area (9) includes at least a portion of said supply means (1), said artificial vision system (3) for identifying and counting items passing through said viewing area (9); a comb (4) for temporary retention of articles driven by said feeding means (1), said comb (4) being placed downstream of said viewing zone (9) and comprising a plurality of teeth (10, 11); 31, 33) which can each take two positions, namely a rest position and a retention position, wherein said tooth does not hinder and respectively impede the passage of articles; said artificial vision system (3) delivering a control information (14) for the comb (4) so that the comb teeth are lowered from the rest position to the retention position until the end of the conditioning cycle. during which the lowering of certain teeth is delayed so as to selectively allow the passage of the last article or articles (32) making it possible to add the number (N) of articles to said predetermined quantity   (No) for the current conditioning cycle.   2. Machine according to claim 1, characterized in that it comprises a barrier (5) for temporary retention of articles driven by said supply means (1), said barrier (5) being placed downstream of said retention comb temporary (4), the lowering of said barrier (5) until the end of the current conditioning cycle being delayed with respect to the lowering of the comb (4), so as to allow the passage of said one or more said articles (32) for refilling, lowering said barrier (5)   authorizing the raising of said comb (4) before the end of the current conditioning cycle.   3. Machine according to any one of claims 1 and 2, characterized in that   it comprises, upstream of said zone of vision (9), a brush (12) placed   above said feed means (1) and rotating in the direction opposite to the driving direction of said feed means, so as to eliminate any overlap   articles on said supply means.   4. Machine according to any one of claims 1 to 3, characterized in that   said feed means (1) comprise, upstream of said viewing zone (9), a   acceleration zone (13), for spreading said articles on said supply means.   5. Machine according to any one of claims 1 to 4, characterized in that   said camera (7) is a matrix camera.   6. A method of sorting and grouping individual articles in view of their conditioning by predetermined quantity during successive conditioning cycles, said articles (2) being driven to a conditioning device by feed means (1), characterized in that it comprises, for each conditioning cycle, the following steps: - tracking and counting (41), by an artificial vision system (3) including in particular a camera (7) whose viewing area (9) encompasses at least a portion of said feed means (1), articles passing through said viewing zone; and, when said predetermined quantity (No) of articles is reached for said conditioning cycle: - calculation (43), by said artificial vision system (3), of information (14) for controlling a comb ( 4) for temporary retention of articles driven by said feeding means (1), said comb (4) being placed downstream of said viewing zone (9) and comprising a plurality of teeth each of which can take two positions, namely a rest position and a retention position, in which said tooth does not hinder and hinders respectively the passage of the articles, said information (14) for controlling the comb being such that the lowering of certain teeth is delayed so as to allow selectively passing the last article or articles (32) for adding the number of articles (N) to said predetermined quantity (No); - lowering (44) of the teeth of the comb (4), from the rest position to the retention position, until the end of the current conditioning cycle, according to said information (14) for controlling the comb. 7. Method according to claim 6, characterized in that said steps (41 to 43) implemented in said artificial vision system (3) for each conditioning cycle constitute an iterative process, each iteration is chained to the previous one and comprises said step (41) of identifying and counting articles passing in the viewing zone and, only when said predetermined quantity (No) of articles is reached for the current conditioning cycle:   said step (43) for calculating comb control information.   8. Method according to claim 7, characterized in that said step (41) for locating and counting articles passing in the viewing zone (9) comprises the following steps: storing (51) a digital image taken by said camera (7); calculating (53) the number of new items that have appeared in the viewing area (9) by comparing the last stored picture with the previous stored picture; calculating (54) the cumulative number of articles passed in the viewing zone (9) since the beginning of the current conditioning cycle; - comparing (42) said cumulative number of articles with said quantity predetermined number of articles.   9. Method according to claim 8, characterized in that said step (43) for calculating a comb control information comprises the following steps: searching (61) groups of articles on said last stored image, each group (82) comprising one or more adjacent articles and constituting an indivisible assembly by said comb (4); calculating (62), according to a predetermined strategy, a selection of at least one group and a lowering order of the associated teeth selectively permitting the passage of said group of the selection, the item or articles of said group of the selection constituting said one or more last articles (32) for adding the number (N) of articles to the quantity predetermined (No).   10. Method according to any one of claims 8 and 9, characterized in that   said step (53) for calculating the number of new items that have appeared in the viewing area (9) comprises the following steps: - searching (71) tasks on said last stored image, each task (81) comprising one or more adjacent items and being viewed by said artificial vision system (3) as having a single contour; determining (72) the new tasks that have appeared in the viewing zone (9), by comparing the tasks of the last stored image with those of the previous stored image; calculating (73) the number of new articles contained in said new tasks, the number of new articles contained in each new task being obtained by dividing the surface of said task by   a predetermined article surface.   Method according to claims 9 and 10, characterized in that each group (82)   consists of a whole number of tasks (81).   12. Method according to any one of claims 6 to 11, characterized in that   comprises, after said step (44) of lowering the teeth of the comb (4) and when said predetermined quantity (No) of articles is reached for the current conditioning cycle, an additional step (45) of lowering of the teeth of the comb (4) a barrier (5) for temporary retention of articles driven by the feed means (1), said barrier (5) being placed downstream of said temporary retention comb (4), lowering said barrier (5) until the end of the current conditioning cycle being delayed with respect to the lowering of the comb, so as to allow the passage of said one or more last articles (32) making it possible to top-up, lowering said barrier ( 5) authorizing the   raising said comb (4) before the end of the current conditioning cycle.