WO2007068697A2

WO2007068697A2 - Apparatus and method for sorting objects

Info

Publication number: WO2007068697A2
Application number: PCT/EP2006/069605
Authority: WO
Inventors: Marco Tartagni; Alessandro Golfarelli
Original assignee: Sortron S.R.L.
Priority date: 2005-12-13
Filing date: 2006-12-12
Publication date: 2007-06-21
Also published as: ITFO20050010A1; WO2007068697A3

Abstract

The present invention relates to an apparatus and methods for classifying and sorting small- and medium-sized objects (from 1 mm to various cm) using electronic systems and advanced sensors. The objects to be sorted are immersed in a continuous flow along guides made parallel on a flat support. One or more electronic systems arranged in predetermined positions along each guide line, without interfering with the flow, classify the single objects on the basis of their physical and/or geometric features. Various electronic systems concur to processing the classification of the sorted flow of objects in real time singularly or collectively, by correlating date from different sensorial elements. The results of the classification performed on the objects moving along each guide line are used to activate sorting systems integral with the lines themselves.

Description

APPARATUS AND METHOD FOR SORTING OBJECTS

DESCRIPTION

Prior Art

The first modern sorting machine (early 1930's) consisted of a moving conveyor belt on which grains of rice to be sorted where placed. The operator could adjust the feed speed of the belt and was also equipped with a flexible tube coupled to a centrifuge pump connected to a sorting cyclone. The operator manually moved the tube towards the non-compliant grains, thus eliminating them by suction and sorting in the cyclone. Such machine, in addition to being labour- intensive, does not ensure high or reliable sorting standards in time. The progress of consumptions spurred the need to improve sorting techniques.

Additionally, sorting needs evolved during the years, to include the sorting of cereal grains, in particular of durum wheat. Such demand is determined by the need of durum wheat flour production companies to standardise the colour of the extruded pasta and increase its organoleptic features.

The first sorting machines based on photoelectric devices allowing full automation and sorting control were made during the 1980's mainly by Sortex. US patents US 4,513,868; US 4,630,736; US 4,699,273; and US 5,873,470 are the most significant examples of intellectual property developed in this field.

The machine was devised so as to convey the grains one by one to a slide which let them drop onto a lighting system. The light reflected by the grain was focused by means of an optical assembly, consisting of filters and focusing lenses, onto a photocell assembly. The signals output by the photocells were electronically compared against a sample signal representing a standard grain. When an anomalous signal was found, the electronic system opened a solenoid valve which sent a jet of compressed air that deviated the grain onto a conveying channel communicating with a collection silos. Such technology was limited in terms of detection accuracy and sorting speed.

In the course of year 2000, new sorting technologies based on optical detectors whose images were processed by microcomputers were introduced. Such machines were based on the use of particularly sensitive multiple colour TV cameras which allow to detect defects even smaller than 0.3 mm. The TV cameras replace the use of photocells allowing sorting nine times more refined and faster than that of the earlier photocell machines. The mechanics used to convey grains to the viewing and non-compliant ejecting assembly remained essentially identical to that of the photocell machine.

In the optical sorting machine, the grains were illuminated by a light having an appropriate wavelength and then viewed by a TV camera provided with a processor which analysed the image and sent, if required, the signal to a pneumatic solenoid valves to eject the non-compliant groups of grains. Such a machine can currently identify grains with black spots, decolouration, unripe green grains, yellow and burnt grains, incompletely dehulled grains, gravel and glass fragments.

US Patents 6,427,128 and 5,917,927 by Satake Corporation are some examples of intellectual property developed in this field.

Summary of the Invention

The invention to which this patent is related aims at seeking faster speed and lower cost methods for characterising granular objects in general, and cereals or grains or seeds in particular. The invention is based on recent electronic technology developments and printed circuit technology allowing to compete with the current technologies and open an investigation niche which is unexplored to date in the state of the art of this field.

Indeed, recent experimental tests conducted by the present applicant (Sortron S.r.l.) have shown how impedimetric analysis performed singularly on single objects may open new horizons to the classification of objects obtaining very marked results in object classification. In particular, impedimetric analysis, as shown in the details of the invention, allows to relate conformational features of the single object, such as volume and shape, to the signal generated by said impedimetric measurements. Finally, the combined and crossed analysis of optical measurements and impedimetric measurements on the single object in a continuous flow have proven to exceed in sorting quality (in particular for grains) those of the state of the art. The main advantages of the present invention consist in the very high discriminatory capacity and the high degree of parallelism, capable of processing large amounts of material on production lines. The use of integrated electronic systems also allows high operative speed, easy calibration and flexibility to the required classification features. Summarising, the advantages of the approach, in the specific case of grain sorting, are: 1) Sorting: the sorting machine allows to obtain, thanks to the previously described crossed analysis process, a product in which the different typologies of objects (specifically, grains) are discriminated and differentiated with very low percentages of error with respect to the sorting output of competing machines currently used in the milling industry; 2) Flexibility: the possibility of using variable selective parameters allows to adapt sorting criteria to the grain batch under investigation; the sorting algorithm which manages the discrimination process and the various types of grains, and therefore the machine output, is actually programmable via software in an entirely flexible manner; 3) Rejected product recovery: three-dimensional characterisation of each grain offered by the system allows a rather efficient use in the recovery of good grain from a rejected product than other sorting machines; and

4) Competitive costs: the structural components of the proposed structure can be made by exploiting widely used materials (the external structure is metallic, the internal piping system is made of plastic). The electronic system itself uses very low cost components and exploits standard PCB (printed circuit board) technology.

Brief description of the drawings A detailed description of a preferred embodiment of the apparatus and method for sorting objects of the present invention will be given hereinafter, as a non limiting example, and with reference to the annexed drawings, wherein: Figure 1 shows a device with guides for conveying the flow of objects with integrated electronic systems;

Figure 2 shows the device of Figure 1 in frontal/cross-sectional view, wherein there are shown an electromagnetic sensor consisting of electrodes P1 and P2, an electronic signal generation system ES, and an optical sensor consisting of an emitter L and a receiver IR;

Figure 3 is a frontal/cross-sectional view of the device of Figure 1 showing the interaction with the electromagnetic field EF generated by electrodes P and the possibility of moving objects by using a conveyor CR; Figure 4 is a top view of the device of Figure 1 showing the interactions with the electromagnetic field EF generated by electrodes P with objects of different conformational structure;

Figure 5 shows the device in Figure 1 in a side view thereof;

Figure 6 shows a functional embodiment of an apparatus for sorting objects according to the present invention;

Figure 7 shows a schematic representation of the circuit control of each electronic system (ES);

Figure 8 shows an electronic oscillator controlled by the electric capacitance established between the metallic electrodes P1 and P2 positioned near guides of the apparatus according to the present invention;

Figure 9 shows a light radiation receiver reading circuit;

Figure 10 shows a controlling system for a multi-channel apparatus wherein each channel includes several electronic system and interacting with the ejection systems (EJ) and the management systems (MC, CH) of the apparatus according to the present invention; and

Figure 11 shows an example of statistic classification of various subcategories of objects detected according to the output of optical sensors (left diagram) and impedimetric sensors (right diagram).

Detailed description of the Invention

With reference now to Figure 1 and as claimed in the present application the invention is based on a multi-sensorial characterisation, and more specifically on crossed optical and impedimetric analysis of the objects to be sorted. Parallel guides (W), also called channels, are created on a slanted plant support (B).

The objects to be sorted are immersed in a continuous, free-falling flow along said guides. The support may be made using generic printed circuit board (PCB) technology which allows to obtain structures with excellent mechanical features, and the possibility of integrating electronic circuits and cost-effective production. A flow of objects (S) proceeds in direction (F) without obstacles of mechanical origin, possibly and preferably in a reciprocally disjoined fashion. To help this condition, a mechanical apparatus (M) facilitates introduction and defines the falling frequency of single objects.

One or more electronic systems (ES) arranged at predetermined positions on each guide line, without interfering with the descending flow, classify the single objects on the basis of their physical and/or geometric features. The electronic systems may be made on the same sliding support (B) or on separate bases containing the single electronic components (EC) and connected to the base by means of specific adapters (A).

With reference now to Figure 2, the presence of two sensorial typologies is highlighted, whose data will be processed in real-time by cross-correlation. In particular, the object is optically analysed by an optical radiation emitter (L) and by a receiver (IR) arranged near the guide line. Without limitations of generality, the emitter may be positioned either on the opposite side of the receiver (as shown in Figure 2) for transmission analysis; or on the same side as the emitter by reflecting analysis of the radiation. The receiver may be positioned in a compartment underneath the guide (W) and inserted in an opening (H) which allows the reception of the light signal. In addition to the above, the two optical detection systems (for transmission and for reflection) may be combined in a structure having one emitter and two receivers, one of which is placed on the same side as the emitter and one is positioned on the opposite side.

In all of the described structures, in order to avoid interference with the surrounding environment, a specific shielding (C) may be placed near the receiver and possibly also supporting the transmitter. Optical analysis is made by an electronic circuit (IRC in Figure 11) capable of detecting and amplifying the signal from the receiver. Entirely independently from above, one or more impedimetric type analysis will be performed on the descending objects. The single object is subjected to a superficial electronic field generated by one or more electrodes (P) connected to electronic circuits (EC) and arranged near the guides. The relationship between voltage applied to the electrodes and respective induced current determines a characterisation of the object (S) susceptible to classification.

The arrangement of electrodes may be made following different typologies to optimise performance and to adapt to the structures of the objects themselves. In this case, the electrodes, made by simple PCB technology metallization may be extended towards the deeper part of the cavity with conductive paints or may remain on the surface.

A simple, but not exclusive, methodology for determining impedimetric features of the single object are shown in Figure 2 and Figure 8. The above- mentioned electrodes constitute the armatures of an electrical capacitor which constitutes the characteristic reactance of an oscillator (OC). Said armatures form an electrical field which intersects the guides. As better shown in Figure 4, the passage of the single object in said guides determines a perturbation of said electrical field inducing a variation of the aforementioned electric capacitance and the characteristic frequency of said oscillator. The oscillator frequency is therefore the index of some conformational features of the grain itself, such as volume, material features and shape.

With reference to Figure 1 , it may be observed how different sensorial analysis types may be adopted: one or more impedimetric and/or optical sensors may be arranged or intercalated along the same guide line in order to collect the highest possible number of data. In a more compact arrangement, the impedimetric and optical sensor may share the same location. The electronic system (EC) may therefore consist of several sensorial units connected to a central unit which may consists of a micro-controller (LC) as shown in Figure 7. Furthermore, for sorting needs, the same unit may control the ejector (EJ) through an actuating circuit (AS). The sorting system uses either inter-channel (between different channels) or intra-channel (within the same channel) data processing and programming. In the first case and as shown in Figure 10, the programming of the single units, connected together by an internal bus (IB) may occur by activating a central unit (CH) and a peripheral controller (MC). In the second case, the data from the sensors arranged on the same channel, processing in real time during descent, are analysed and processed locally by smart units (LC) which will evaluate the actuations on the basis of discrimination thresholds (TH) as shown in Figure 11.

As concerns the flow of elements to be classified/sorted, alternatively to the approach shown in Figure 2 where such objects are immersed in the guides in a continuous free-falling flow, a different methodology may be used as shown in Figure 3 (and Figure 4 right channel), wherein a mobile conveyor system (CR) positioned within each guide line is responsible for conveying the flow of objects only while the electronic devices adapted to measure and characterise the single elements remain integral with the guide itself. By exploiting the cross-correlation process of the information generated by the two sensorial typologies used, the statistic distributions related to the optical and impedimetric information characterising each type of object under examination may be identified.

In the specific case of Figure 11, the optical and impedimetric distributions are related to the different types of seeds which constitute durum wheat. These characteristic distributions result from a rather large number of samples of each type of object to be classified and are used as comparison curve to identify which family of objects each element in the flow to be sorted belongs to.

Discrimination thresholds (TH) are defined on the sample curves of the type shown in Figure 11 which identify ranges (AR) in which there are different statistic probabilities of finding the same types of objects. By combining the ranges identified in the graphs related to the optical information and those related to impedimetric information, according to an appropriate sorting algorithm, all types of objects to be sorted can be identified with a very low likelihood of error.

Therefore, the analysis method according to the present invention envisages a first self-learning step of the apparatus: during this step, the characteristic sample distributions are extrapolated and associated to each typology of object to be classified/sorted and consequently an appropriate selection algorithm is defined on the basis of the set thresholds (TH) and the resulting ranges (AR). The second step consists in the classification itself, where each object to be identified is detected by the two sensorial typologies and the information associated thereto is compared against the sample distributions in order to identify the typology to which it belongs. The results of the classification made on the continuous flow along each guide line are used to activate sorting system integral with the lines themselves.

Object sorting, which may occur in the forms explained below, is generically called electromechanical sorting.

A first method of electromechanical sorting of the objects may be of pneumatic or mechanical nature directly on the base which accommodates the flow (B). In one case, the objects considered non-compliant to specifications (or vice versa) will be ejected using pneumatic systems (EJ) which will induce the ejection of gas from nozzles. In another possible embodiment, the ejection will be performed by mechanical action of elements which are activated by electronically controlled servo mechanisms.

An alternative method for sorting objects envisages the action of an electrical field on the objects once their flow has completed to travel the entire base (B) and therefore they are either free-falling or however beyond the area of analysis. In practice, the objects, electrostatically pre-charged before or during their flow in the base (B), will be then deviated along their course and therefore ejected by means of the action of a strong electrical field .

Claims

1. An apparatus for classifying a plurality of objects (S) comprising:

- devices for conveying said objects (S) in a controlled flow having a direction (F) so that said objects are moved and separately conveyed in said direction (F) of said controlled flow;

- guides (W) longitudinally arranged on a flat support (B) and for guiding said objects (S) in said controlled flow; and

- electronic systems (ES) arranged in proximity of one or more positions of said guides (W) and for individually characterising said objects (S) without interfering with said controlled conveying flow.

2. An apparatus for sorting a plurality of objects (S) according to the preceding claim, further comprising electromechanical devices (EJ) controlled by said electronic systems (ES) and for separating subcategories of said objects (S) from said controlled flow by means of individual characterisation.

3. An apparatus according to claims 2, wherein said electromechanical devices (EJ) comprise electro-pneumatic systems for separating by gaseous emissions or mechanical actions subcategories of said objects (S) from said controlled flow by means of individual characterisation.

4. An apparatus according to claims 2 or 3, wherein said electromechanical devices (EJ) comprise electrostatic acting systems for generating electromagnetic fields so as to separate subcategories of said objects (S) from said controlled flow by means of individual characterisation.

5. An apparatus according to anyone of the preceding claims 1 to 4, wherein said conveying devices comprise introduction mechanical systems (M) for conveying said controlled flow along said guides (W) on said flat support (B).

6. An apparatus according to anyone of the preceding claims 1 to 4, wherein said conveying devices comprise mobile feeding systems (CR) aligned to said guides (W) and for guiding said controlled conveying flow.

7. An apparatus according to anyone of the preceding claims 1 to 6, wherein said electronic systems (ES) comprise electronic oscillators (OC) which comprise: - one or more electrodes (P) arranged in proximity of one or more positions of said guides (W) and for generating an electromagnetic field (EF) which individually interacts with said objects (S) in said controlled flow; and

- electronic components (EC) electrically connected to said electrodes (P) and for generating electromagnetic oscillations whose characteristic frequency depends on the interaction of said electromagnetic field (EF) with each object (S) in the controlled flow.

8. An apparatus according to anyone of the preceding claims 1 to 6, wherein said electronic systems (ES) comprise optical detection devices (IRC), which comprise:

- generating means (IE) of said optical radiation (LIG) for interacting with said each single objects (S) in said controlled flow;

- sensing means (IR) of said optical radiation (LIG) for interacting with said single objects (S) in said controlled flow; and - electronic components (EC) electrically connected to said emission means

(IE) of said optical radiation (LIG) and to said sensing means (IR) for measuring said interaction of said optical radiation with said objects (S) in said controlled flow.

9. An apparatus according to claim 8, wherein said sensing means (IR) are arranged to detect said optical radiation (LIG) transmitted through said single objects (S) in said controlled flow.

10. An apparatus according to claim 8, wherein said sensing means (IR) are arranged to detect said optical radiation (LIG) reflected on said single objects (S) in said controlled flow.

11. An apparatus according to claim 8, wherein said sensing means (IR) concurrently detect said optical radiation (LIG) transmitted on said single objects

(S) and said optical radiation (LG) reflected on said single objects (S) within said controlled flow.

12. A method for classifying a plurality of objects (S), characterised in that it comprises the following steps: - conveying said objects (S) along guides (W) longitudinally arranged on a flat support (B) and along a direction (F) so that said objects are controlled and separately conveyed in said direction (F) in a controlled flowing manner; making said objects (S) moving in a controlled flowing manner and interact with one or more electronic systems (ES) arranged near said guides (W) and adapted to determine features of said objects (S) without interfering with said controlled flow; and storing and/or processing in a computerised system (CH) the information from said one or more electronic systems (ES) and characterising said single objects (S).

13. A method for sorting a plurality of objects (S) according to claim 12, further comprising the step of: activating electromechanical sorting means (RJ) controlled by said electronic systems (ES) for deviating from said controlled flowing manner and by means of gaseous emissions or mechanical actions, subcategories of said objects having similar characteristics.

14. A method for sorting a plurality of objects (S) according to claims 12 or 13, further comprising the step of inducing electrical charges in said plurality of objects before the said step of conveying said objects (S) along guides (W) longitudinally arranged on a flat support (B) in a controlled flowing manner along a direction (F).

15. A method according to anyone of claims 12 to 14, wherein said interaction of said objects (S) with said electronic system (ES) is obtained by means of an electromagnetic field (EF) generated by electrodes (P) electrically connected to said electronic systems.

16. A method according to anyone of claims 12 to 14, wherein said interaction of said objects (S) with said electronic systems (ES) is obtained by means for emitting and/or reading optical radiation (LIG) transmitted and/or reflected from said objects (S) flowing in a controlled manner.

17. A method according to anyone of claims 12 to 14, wherein the classification of single objects (S) is obtained by a further step wherein is accumulated information concerning the same single object (S) and coming from several electronic systems (ES) associated to a specific guide (W) in a computerised system for increasing the characterisation and/or discrimination capacity.

18. A method for classifying a plurality of objects, characterised in that it comprises the following steps: - conveying sample objects belonging to one or more similitude categories along guides (W) longitudinally arranged on a flat support

(B) in a controlled flowing manner and along a direction (F) so that said sample objects are controlled and separately conveyed along said direction (F); - making said sample objects moving in said controlled flow in manner interact with electronic systems (ES) arranged near said guides (W) and for determining the features of said sample objects without interfering with said controlled flow; accumulating information from one or more electronic systems (ES) related to said categories of similitude in a computerised system so as to determine the statistic distribution (G) of the features of said categories of similitude; identifying decision thresholds (TH) characterised by said statistic distributions; - conveying objects to be classified along guides (W) longitudinally arranged on a flat support (B) in a controlled flowing manner and along a direction (F) so that said objects to be classified are controlled and separately conveyed in said direction (F); making said objects to be classified interact with electronic systems (ES) arranged near said guides (W) and adapted to determine the features of said objects to be classified without interfering with said controlled flow; and accumulating in said computerised system the information related to said objects to be classified so as to determine affinities of each object to be classified with said sample object categories.

19. A method for sorting a plurality of objects according to the preceding claim, further comprising the following steps: comparing the information from one or more electronic systems (ES) and related to said objects to the be classified according to said decision thresholds; and activating electromechanical sorting means (EJ) on the basis of comparison of said information related to said objects to be classified according said decision thresholds for deviating subcategories of objects having features of affinity to said sample object categories from said controlled flowing manner of objects to be classified.