GB2060166A - Sorting device and method - Google Patents

Abstract

An infrared light generator directs a flashing band of light across a conveyor belt carrying randomly spaced materials having differing properties. The frequency of the flashing band sequentially alternates between two infrared frequencies chosen for their peculiar reflectivity characteristics. Infrared light sensitive cells detect energy reflected from the passing materials as they are exposed to the infrared light. Comparator circuitry examines the detected information, determining the size and material nature of the passing pieces. Powered paddles act at the direction of the comparator, physically separating variously sized inorganic from organic materials, or ripe from unripe comestibles. <IMAGE>

Description

SPECIFICATION Sorting device and method The present invention relates to a process and apparatus for sorting materials.

Mechanical harvesters capable of uprooting and then denuding plants and their comestibles have vastly improved the speed and overall efficiency of the harvesting process. However, a major problem still attendant upon mechanical harvesting includes the initial sorting of comestibles from field debris and dirt clods. On the earliest harvesters, this primary sorting job was performed by a worker who visually examined the passing mixture of produce and debris and removed acceptable comestibles for further sorting. Manual sorting is not only labor consuming but also wastes produce that is discharged onto the ground before it can be manually removed from the transport conveyer.

When electronic sorters were first developed, colorimetry emerged as the most reliable and effective basis upon which to sort objects. The earliest colorimetry systems were monochromatic, but the later, more sophisticated systems used a bichromatic approach. One bichromatic system used a source of constant light energy to bathe the subject materials. Then, dual detectors, each fitted with an appropriate filter, gather information regarding the reflected light The dual detectors are interconnected to comparator circuitry which produces a reject signal if the input signals bear a predetermined relationship. The reject signal is fed to a powered ejector arm which deflects the unwanted object from the stream of articles.

The advantages of a bichromatic colorimetry sorter are manifest. The approach can be used for differentiating ripe from unripe comestibles, or organic from inorganic materials. Equally important, the signal to noise ratio afforded by the bichromatic system provides superior quality data for logic circuits to analyze than the monochromatic system, thus greatly increasing the reliability of the sorter apparatus.

Prior art bichromatic sorters do not perform well in separating organic from inorganic materials because the detectors used, generally photovoltaic silicon cells, are not sufficiently sensitive to the deep infrared light which inorganic materials reflect. Lead sulfide cells, however, perform well as deep infrared detectors, but have not been used successfully in prior art devices.

Changes in temperature drastically affect the resistance of the photoresistive lead sulfide cell.

Variations between the temperature sensitivity of individual cells result in different resistance values for a group of cells throughout a temperature spectrum. With the dual detector approach used in the prior art, temperature variations affect the relative resistance of two companion cells at differential rates. A relatively small change in ambient temperature demands recalibration of both detector circuits.

The requirement of frequent recalibration proved too time consuming and technically difficult to perform in the field. But without frequent recalibration, temperature variations encountered during in-field use of a dual detector sorter using lead sulfide detectors rendered the sorter unreliable and commercially unusable.

Thus, the inability of prior art sorters to use the superior lead sulfide cell successfully stems from the use of two detectors to extract data from the reflected light.

The present invention, by changing the nature of the source of illumination, uses only one cell for detecting reflected light. Temperature variations do not degrade the reliability of the single detector cell system since the sensitivity of the material analysis circuitry can readily be adjusted, if necessary, to compensate for temperaturedependent sensitivity variations in a single lead sulfide detector cell. Furthermore, for ripe/unripe sorting, the single detector cell system can be used to advantage with the lower sensitivity, silicon detector cell as well.

The present invention provides a process for sorting two materials, each material exhibiting an amplitude peak in reflected light wave energy at a different characteristic frequency, comprising the steps of: exposing a planar area defining a sampling zone to alternating pulses of light wave energy at two distinct frequencies, each of said pulses corresponding in frequency, respectively, to the characteristic frequency of amplitude peak in reflected light energy for each of the materials; moving the materials through the sampling zone; detecting the alternating pulses of light reflected from the materials moving through the sampling zone; producing a reject signal if the detected pulses bear a predetermined relationship; and physically separating one of the materials from the other material in response to the reject signal.

The present invention also provides an apparatus for sorting two materials, each material exhibiting an amplitude peak in reflected light wave energy at a different characteristic frequency, comprising: means for exposing a planar area defining a sampling zone to alternating pulses of light wave energy at two distinct frequencies, each of said pulses corresponding in frequency, respectively, to the characteristic frequency of amplitude peak in reflected light energy for each of the materials; means for moving the materials through the sampling zone; means for detecting the alternating pulses of light reflected from the materials moving through the sampling zone; means for producing a reject signal if the detected pulses bear a predetermined relationship; and means for physically separating one of the materials from the other material in response to the reject signal.

The invention will be best understood from the following description of the accompanying drawings, in which: Figure 1 is a schematic representation of the physical components of the invention shown interconnected to a block diagram of the associated electrical circuitry; Figure 2 is a graphic representation of the reflectivity for both dirt and tomatoes exposed to a plurality of test frequencies with a specific infrared light energy spectrum, Figure 3 is a side elevational view of the rotatable drum, the major element of the light splitting apparatus; Figure 4 is a vertical median cross sectional view, taken on the plane indicated by the line 4-4 in Figure 3, illustrating the circumferentially positioned longitudinal apertures and adjacent twin filters with depending bar lens;; Figure 5 is a fragmentary cross sectional view taken as in Figure 4, but to an enlarged scale and showing both filters shielded from the illumination source; Figure 6 is an electrical schematic of twin preamp channels used to amplify the infrared detector output; Figure 7 is an electrical schematic of the pulse generator circuitry, showing the tach generator, frequency counter with decoded output, and reset generator; Figure 8 is an electrical schematic of twin comparator channels showing the unity gain followers, quad comparators, sizing circuits, delay and stretch circuitry, and valve driver components; Figure 9 is a graphic representation of reject signals of an undersized object and a minimum sized object as they pass through selected points in the sizing circuitry; Figure 10 is a graphic representation of the effect the delay and stretch circuitry has on a reject signal; and Figure 11 is a schematic of the low voltage power supply in which the V+ and V symbols are respectively connected to all V+ and V symbols in the other figures.

Since the immediate commercial advantages offered by the present invention are a process and apparatus capable of effectively physically separating inorganic from organic materials, the embodiment offered herein will describe in detail an inorganic-organic sorter. In a manner explained below, however, the process and apparatus are readily adaptable for sorting unripe from ripe comestibles, and it is to be understood that the present invention embraces both embodiments as well as variations thereof apparent to one skilled in the art.

The light frequencies used in a bichromatic colorimetry system for sorting inorganic from organic materials are derived from the natural reflectively characteristics of the materials to be sorted. Making reference to Figure 2, reflected light energy, expressed as a percentage of incident light, is plotted against frequency, expressed as wave lengths in microns, for the two subject materials. Tomatoes exhibit a high, sharp, energy reflectance peak at 0.8 microns while dirt shows a moderately high, but broad peak at 1.8 microns.

Fortunately, the reflectance of dirt at 0.8 microns is relatively low and the reflectance of tomatoes at 1.8 microns is quite low, indeed. It is this relative difference in reflectivity for tomatoes and dirt at two separate frequencies which provides the basis for bichromatic colorimetric sorting pursuant to the present invention.

The present invention uses a dual frequency illuminator to produce sequentially ordered bursts of infrared light energy at 0.8 and 1.8 microns, for inorganic/organic sorting. The dual frequency illuminator 11 is shown generally in Figure 1 and Figures 3 and 4 reveal specifics of its key element, a rotatable drum assembly 12 including a horizontaily-disposed, hollow, right circular cylinder 13 having a plurality of circumferentially positioned, elongated apertures 14. Drum hubs 15 and spinder arms 1 6 support and journal the cylinder 13 for rotation about axle 1 7. The lefthanded extremity of axle 17, as seen in Figure 3, is mounted on left-hand support bracket 18.Righthand support bracket 1 9 provides a bearing for drum hub extension 21, which protrudes beyond the right-hand extremity of axle 1 7. Drive pulley 22 is attached to the right-hand end of sleeve extension 21 so that when in operation, cylinder 13 rotates when rotational torque is applied to drive pulley 22, and the axle 17 remains stationary, being constrained by left-hand support bracket 1 8 to which the axle 17 is secured.

A plurality of lamps 23 depends from the underside of fixed axle 1 7, the lamps being directed to illuminate an arcuate section of cylinder 13 immediately therebeneath. An infrared organic filter 14 and an infrared inorganic filter 26 are adjacently co-planar beneath and extend the full length of right cylinder 13 (see Figures 3 and 4). A convex in section bar lens 27 is positioned beneath and is co-extensive with the planar bottom surface of the filters 24 and 26.

As shown in Figure 1, drive motor 28, tach generator 29 and drive pulley 22 are interconnected by belt 31. As the drive motor 28 is energizes, the cylindrical drum 1 3 begins to rotate in a ciockwise fashion as viewed in Figure 1. Light energy generated by the lamps 23 passes first through apertures 14 and then through filters 24 and 26 sequentially before being blocked by cylinder wall section 32.

The transverse dimensions of apertures 14, organic filter 24, and inorganic filter 26 are substantially identical. Cylinder wall section 32 is approximately twice the transverse dimension of aperture 14. This effects a repetitive, sequential pattern of light pulses emanating downwardly through the two infrared filters, each pair of pulses being separated by a blank dark period when cylinder wall section 32 blocks the illumination source, lamps 23.

The filtered pulse emanating downwardly from each filter is collected and directed by bar lens 27 to illuminate a relatively narrow transverse portion of the turnaround of discharge end of a supply conveyer 33. As Figures 3 and 4 reveal, there are 8 groups of apertures 14 equally spaced about the cylinder 13. Each full revolution of cylinder 13, therefore, produces 24 bits of information, i.e., eight cycles with each cycle consisting of a pulse through filter 24 followed by a pulse through filter 26 followed by a dark period.

Organic comestibles 34 and inorganic debris 36, randomly disposed upon the supply conveyor 33, emerge into a sampling zone 35 at the discharge end of the conveyor 33. As illustrated in Figure 1, filtered, sequential pairs of pulses from the dual frequency illuminator 11 impinge upon comestibles 34 and debris 36, and a portion of the reflected energy reaches collector lens 37 within a detector assembly 38. Collector lens 37 focuses the pulses upon a lead-sulfide detector cell 39.

While Figure 1 shows a single lens 37 and cell 39 for purposes of simplicity, actual practice of the invention requires a plurality of lenses and respective cells to account for all articles passing through the sampling zone 35.

Tach generator 29 is an AC generator producing 12 Hz per revolution. The generator 29 and the cylinder 1 3 are belt driven together at identical rates. Thus, the generator 29 and the cylinder 1 3 are in synchronism so that when the sine wave output of the tach generator 29 passes through zero potential, the apertures 14 will be centered over either the fiber optic assembly 41 (as shown in Figure 5), the organic filter 24, or the inorganic filter 26. The import of this synchronous pulse will become clear when the demodulator circuitry is discussed immediately below.

Figure 7 shows the tach generator 29 interconnected to two bridge rectifier circuits 43 and 44. Negative high voltage for biasing the leadsulfide detector cell 39 of pre-amplifier 42 (see Figure 6) is provided at point C by negative supply circuit 43. Pulse supply circuit 44, on the other hand, rectifies the 12 HZ output of the tach generator 29 to produce a 24 HZ pulsating DC voltage. A diode 46 clips the high voltage peaks of the pulsating DC at 10 volts, leaving abbreviated pulse "valleys" of 24 HZ. Wave-shaping NOR Gates 47 further narrow the pulse width before the pulse signal is fed into a counter 48, or synchronous demodulator. Referring now to Figure 5, the cylinder 13 is in a "dark" position since the opaque cylinder wall section 32 is preventing light from reaching either the organic filter 24 of the inorganic filter 26.One of the apertures 14, however, is centered over the fiber optic assembly 41, permitting light to fall upon a fiber optic tube 51. The light energy is carried through the tube 51 to a reset generator 52 (see Figure 7) where it causes photo diode 55 to conduct. The signal, augmented by amplifier 54, is fed into the counter 48. Each time the cylinder 13 reaches a "dark" position, a reset pulse causes the counter 48 to recycle, producing another series of sequential pulses to output AND Gates 53. Each sequential synchronous pulse corresponds either to a "dark" position (output D), or a near infrared pulse through the organic filter 24 (output 0), or a deep infrared pulse through the inorganic filter 26 (output 1).

Figure 8 illustrates two complete channels of comparator, threshold, sizing, delay and stretch, and valve driver circuitry. Two adjacent channels are shown so that commonality and cooperation between elements are illustrated.

With the cylinder 13 in a "dark" position, a synchronous pulse from output 0 of the AND Gates 53 causes a "dark" analog switch 54 to conduct. This prepares capacitors 56 and 57 in Channels A and B, respectively, in Figure 8, for storing the instantaneous signal from their respective detector cell 39 and preamplifier 42. In the "dark" position, the only signal present is the ambient light detected and the resting gain of preamplifier 42. Thus, a "dark" signal value is stored in capacitors 56 and 57 respectively.

When the cylinder 13 rotates in counter clockwise fashion as viewed in Figure 5, apertures 14 leave the fiber optic assembly 41 to illuminate the organic filter 24. A synchronous pulse is produced at output 0 from AND Gates 53 when apertures 14 and the organic filter 24 are in register. This synchronous pulse causes organic analog switches 58 to conduct, and any near infrared light energy reflected from an object passing through the sampling zone 35 is sensed by the detector or detectors 39 focused on that particular area.

Assuming in the first instance, that the object is relatively small and wholly within the field of view encompassed by a single detector 39 feeding channel A, for example, of the preamplifier 42 (see Figure 6), the amplified signal is stored in capacitor 59 in the channel A circuit (see Figure 8). Since the "dark" signal value (ambient + resting gain) stored in the capacitor 56 in the channel A circuit opposes the incoming composite reflected signal value (ambient + resting gain + actual signal), only the resultant, actual signal value is stored in capacitor 59.

With reference now to Figure 4, cylinder 13 continues its rotation to place the longitudinally aligned apertures 14 in register with the inorganic filter 26. A respective synchronous pulse is generated at output I from AND Gates 53 and fed to inorganic analog switches 61. As the organic switches 61 momentarily conduct, an electrical signal corresponding to any far infrared light energy reflected by the object passing through the sampling zone 35 and detected by the detector 39 is stored in the capacitor 62. In a fashion similar to the operation outlined above, the "dark" signal stored in channel A capacitor 56 subtracts from the composite reflected signal value, leaving only the true resultant value of the reflected signal stored in capacitor 62.

Cylinder 13 is rotated at such a rate that the sequential illumination of the sampling zone 35 occurs at approximately 200 times per second. This rate assures that enough information is obtained about the objects passing through the sampling zone so that a material determination can be made. At an identical rate, the reflected sequential pulses are detected and transformed into electrical signals, decoded as discrete bits of information, and finally stored for analysis, as described above.

Thus, the remainder of the detailed description is devoted to a presentation of the information analysis circuitry and cooperating physical separation components.

Figure 8 reveals a plurality of unity gain followers 63 serving to isolate capacitors 62 and 59 from quad comparator 64. Assuming that an object has been sensed by a detector 39 in the channel A circuit and a respective signal level concerning that object has been stored in capacitors 62 and 59, comparator 64 compares the two respective voltages across each capacitor.

If the capacitor 62 voltage is higher than the capacitor 59 voltage, the object is inorganic, and the output ofthe comparator 64 goes high at both output legs of the channel. If the object is organic, the voltage of capacitor 59 is higher than that of capacitor 62, and the output at connection 68 goes low, while output at 69 remains high.

Threshold voltage, introduced at threshold voltage source 70, establishes the lower limit of sensitivity of the comparator 64. If the signal level stored in capacitor 59 does not exceed this minimum threshold voltage, the comparator 64 does not react at all. Thus, spurious or inconsequential signals are eliminated at this point, ensuring reliable operation of the device.

Comparator AND Gate 66 requires a high signal level on both input legs to conduct a signal. Thus, if the object is organic and the voltage at connector 68 goes low, the signal is squelched at that point. If, however, the object is inorganic, a reject signal is passed to sizing circuit 67.

The object to be rejected must be of a minimum size before the reject signal is passed on. Sizing circuitry is necessary to temper the sensitivity of the reject system and assure reliable rejection of only those objects whose size demands their removal from the stream of articles.

With particular reference to Figures 8 and 9, absent a reject signal, voltage at F, G, and H is low.

Clamping diode 75 is on, keeping G low. When a reject signal is passed by comparator AND Gate 66 and appears at F, the output of OR Gate 49 goes high, thereby turning off diode 75 and removing the clamp from point G. The reject signal also passes through a variable resistor 71, serving as a manually adjustable sizing control, and beings charging capacitor 72 at a rate determined by the resultant RC circuit.

The input of sizing comparator 85 is bridged by capacitor 72. Since the lower input leg of comparator 85 is at 5v potential (see Figure 11), the output at H will not go high until the potential at the upper input leg exceeds 5v.

Assuming in the first instance that the object is undersized, capacitor 72 will not reach 5v before the reject signal at F disappears and diode 75 clamps G and thus the capacitor 72 to ground. As Figure 9 shows, the reject signal of the undersized inorganic object is eliminated and there is no output at H.

If the inorganic object is of minimum size, the reject signal continues to charge capacitor 72 past 5v, resulting in an output at H, as seen in Figure 9.

When points F and H are high, sizing AND Gate 50 turns on a diode 80 clamping the capacitor 72 to 1 Ov, the steady value of the reject signal. When the reject signal at F disappears, the diode 80 is turned off and the RC circuit starts discharging towards zero potential. When the voltage across capacitor 72 drops below 5v, the output at H drops to zero, turning on diode 75 and thus clamping the capacitor 72 to ground potential.

The sizing circuit thereby eliminates the reject signal for undersized inorganic objects and passes the reject signal for at least minimum sized inorganic objects.

The delay and stretch circuitry 73 is necessary to account for the time discrepancy between the near instantaneous inorganic/organic determination and the relatively time-consuming and delayed physical separation of inorganic materials. As can be seen most clearly in Figure 1, this delay is the period of time required for the object to travel from the sampling zone 35 to reach the deflection paddle 74. Thus, the reject signal for a particular object must be delayed until that object has just reached deflection paddle 74.

This time-delayed reject signal must also be stretched in duration to compensate for the time lag required for the deflection paddle 74 to respond physically to an electrical signal.

The delay and stretch circuitry 73 comprises a shift register 73 fed by a variable frequency clock 76, or delay oscillator, including a coarse adjustment 77 and a fine adjustment 78. The output frequency of the clock 76 determines the speed and the reject signal passes through the shift register 73. The higher the frequency of clock 76, the shorter the delay of the reject signal through shift register 79.

Stretch of the reject signal is accomplished by feeding two delayed outputs of the shift register 79 into a stretch OR Gate 90. Figure 10 shows the original reject signal H, the 75% signal at J, the 100% signal at K, and the resultant 125% output at L. OR Gate 90 simply adds the two signals J and K to stretch the delayed reject signal to 125% of its original iength.

The reject signal from the shift register 79 is applied to a valve driver 81 circuitry comprising transistors 82, 83 and 84. As premised earlier, the object under consideration is not large enough to be detected by two detector channels. Therefore, only a single reject signal is available, and it is fed to transistors 82 and 83. Diode 86 prevents transistor 84 from being turned on by the adjacent channel's reject signal.

The emitters of transistors 82, 83 and 84 are connected in parallel to one leg of a power supply 87. Outputs X, Y and Z of transistors 82, 83 and 84, respectively, are fed to the coils of individual solenoid-actuated air valves 88 (see Figures 1 and 8) which are, in turn, connected to the other leg of the power supply 87. Each air valve 88 powers a respective air cylinder 89 and interconnected deflection paddle 74, as appears most clearly in Figure 1.

Deflection paddle 74 is held by air cylinder 89 in a normally extended, inclined position as shown in solid line in Figure 1. Until a rejection signal causes the solenoid valve 88 to redirect pressurized air from air supply 91, comestibles 34 continue to fall off the supply conveyor 33, and drop against the deflection paddle 74 which deflects them onto return conveyor 92. In response to the reject signal, however, transistors 82 and 83 turn on, activating their respective solenoid air valves. When support for the deflection paddles 74 is removed, and the impact of the descending debris 36 urges paddles 74 into a position indicated by dotted line in Figure 1. A discharge chute 93 collects the falling debris.

After the reject signal terminates, solenoid air valve 88 substantially instantaneously restores the air cylinder 89, and thus the paddle 74, to its former position.

In the situation where a larger subject article is sensed by two or more detectors 39 and subsequent circuitry determines it to be inorganic in nature, a substantially identical operation occurs. However, rather than a single reject signal appearing, two or more reject signals will be available, depending on how many channels have detected the article. Assuming that both channels A and B have detected an inorganic object, two reject signals would cooperate to turn on transistors 82, 83 and 84. Thus, support for three deflection paddles 74 would be removed and permit the larger piece of debris to pass downwardly through the discharge chute 93, followed by a return of all three paddles to deflecting position.

If ripe/unripe sorting is to be performed, the disclosed device and method for sorting are equally proficient. Depending upon the reflectivity characteristics of the ripe and unripe comestibles to be sorted, filters 24 and 26 will be modified accordingly. For example, if tomatoes are to be sorted, filter 24 should pass red light and filter 26 should pass green light. The basis for bichromatic sorting is the differential in amplitudes of light reflected from red and green tomatoes exposed to alternating pulses of red and green light.

It has been determined, experimentally, that a pulse having a wavelength in the range of 0.65 to 0.68 microns produces a usable amplitude peak for a ripe tomato, and a similar amplitude peak for an unripe tomato has been noted using an exposure pulse having a wavelength in the range of 0.50 to 0.55 microns. While exposure pulses in these general ranges will work suitably for visible light color sorting of many comestibles, precise correlation of ripe/unripe reflectivity constants for the comestible involved with the filters actually used will produce the best results.

Silicon cells, which are quite sensitive to the visible spectrum used for ripe/unripe sorting, can replace the lead sulfide cells used for inorganic/organic sorting. Thus, for ripe/unripe sorting, the detector 39 can be a silicon cell.

Through the simple expedient of providing proper filters and an appropriate detector for the sorting frequencies used, the invention is completely adaptable for ripe/unripe sorting.

Claims (17)

1. A process for sorting two materials, each material exhibiting an amplitude peak in reflected light wave energy at a different characteristic frequency, comprising the steps of: exposing a planar area defining a sampling zone to alternating pulses of light wave energy at two distinct frequencies, each of said pulses corresponding in frequency, respectively, to the characteristic frequency of amplitude peak in reflected light energy for each of the materials; moving the materials through the sampling zone; detecting the alternating pulses of light reflected from the materials moving through the sampling zone; producing a reject signal if the detected pulses bear a predetermined relationship; and physically separating one of the materials from the other material in response to the reject signal.

2. A process as in Claim 1, wherein one of the materials is organic and the other material is inorganic and wherein the alternating pulses of light wave energy are at two distinct frequencies, the first having a wavelength approximately in the range of 0.60 to 0.95 microns and the second having a wavelength approximately in the range of 1.40 to 2.50 microns, respectively.

3. A process as in Claim 1, wherein one of the materials is a ripe comestible and the other material is an unripe comestible and wherein the alternating pulses of light wave energy are at two distinct frequencies, the first having a wavelength approximately in the range of .65 to .68 microns and the second having a wavelength approximately in the range of .50 to .55 microns, respectively.

4. A process as in Claim 1,2 or 3, including the step prior to the separating step of eliminating the reject signal if its duration does not exceed a predetermined period in order to prevent physical separation of an object too small to warrant removal.

5. A process as in Claim 4, wherein the occurrence and stretching the duration of the reject signal is delayed in order to compensate respectively for the physical separation of the sampling zone from the area of physical removal of one material from the other and for the operational time lag in the physical separation step.

6. An apparatus for sorting two materials, each material exhibiting an amplitude peak in reflected light wave energy at a different characteristic frequency, comprising: means for exposing a planar area defining a sampling zone to alternating pulses of light wave energy at two distinct frequencies, each of said pulses corresponding in frequency, respectively, to the characteristic frequency of amplitude peak in reflected light energy for each of the materials; means for moving the materials through the sampling zone; means for detecting the alternating pulses of light reflected from the materials moving through the sampling zone; means for producing a reject signal if the detected pulses bear a predetermined relationship; and means for physically separating one of the materials from the other material in response to the reject signal.

7. An apparatus as in Claim 6, wherein one of the materials is organic and the other of the materials is inorganic and wherein the alternating pulses of light wave energy are at two distinct frequencies, the first having a wavelength approximately in the range of 0.60 to 0.95 microns and the second having a wavelength approximately in the range of 1.40 to 2.50 microns, respectively.

8. An apparatus as in Claim 7, wherein said detection means includes a lead sulfide cell directed toward the sampling zone.

9. An apparatus as in Claim 6, wherein one of the materials is a ripe comestibie and the other of the materials is an unripe comestible and wherein the alternating pulses of light wave energy are at two distinct frequencies, the first having a wavelength approximately in the range of .65 to .68 microns and the second having a wavelength approximately in the range of .50 to .55 microns, respectively.

10. An apparatus as in Claim 9, wherein said detection means includes a silicon cell directed toward the sampling zone.

11. An apparatus as in any of the preceding claims 6 to 10, wherein said means for moving the materials through the sampling zone comprises a supply conveyor belt.

12. An apparatus as in any of the preceding claims 6 to 11, wherein said means for exposing a planar area defining a sampling zone comprises: a frame; a hollow, right circular cylinder mounted on said frame for rotation about an axis parallel to the plane of said sampling zone, the wall of said cylinder including a plurality of circumferentially spaced elongated apertures; means for rotating said cylinder; a fixed light source mounted on said frame and extending through said cylinder, the illumination from said light source being directed from a generally central position along the axis of said cylinder outwardly towards the inner wall of said cylinder in the direction of said sampling zone; a pair of elongated coplanar light filters carried by said frame and positioned side by side adjacent the outer wall of said cylinder and parallel to said planar sampling zone, each of said filters being approximately the same width and iength of said elongated apertures, and affording a characteristic optical frequency corresponding to said characteristic frequency for each of the materials to be sorted; and a lens interposed between said pair of light filters and said sampling zone, said lens being generally coextensive with said pair of light filters so that as said cylinder rotates, a beam of light emerging from each of said apertures and passing first over one filter and then over the other filter, before being obscured by an intervening cylinder wall, is directed toward said sampling zone, the materials moving through said sampling zone being thereby illuminated in each cycle by a pulse of light from one filter, then from the other filter followed by a dark period.

13. An apparatus as in any of the preceding claims 6 to 12, wherein said means for producing a reject signal comprises a comparator having at least two input terminals and an output terminal, said input terminals receiving electrical pulses corresponding to the detected sequential alternating pulses of reflected light, said output terminal producing a reject signal when the first of the electrical pulses, identified with the amplitude peak in reflected light energy for the material to be discarded, is greater in amplitude than the subsequent electrical pulse identified with the material to be retained.

14. An apparatus as in any of preceding claims 6 to 13, including means for eliminating the reject signal if the reject signal does not exceed a predetermined period in duration, said predetermined period being established by reference to the minimum size of the material to be discarded.

1 5. An apparatus as in Claim 11, wherein said means for physically separating one of the materials comprise a return conveyor belt, said return conveyor belt being subjacently co-extensive with and moving in a direction opposite that of said supply conveyor belt; a deflection paddle movable between a first position angularly disposed within the downward path of articles descending from the discharge end of said supply conveyor belt in order to deflect said articles onto the adjacent extremity of said return conveyor belt, and a second position removed from said downward path by the weight of downwardly passing articles striking and displacing said paddle to permit the articles to fall without restriction; means for returning said paddle from said second position to said first position in the absence of a reject signal; and means for deactivating said paddle returning means at the direction of a reject signal.

16. An apparatus as in Claim 15, including electrical means for delaying the reject signal in order to retard the operation of said deactivating means by an amount dependent upon the vertical distance between said sampling zone and said paddle.

17. An apparatus as in Claim 16 including further electrical means for extending the duration of the reject signal in order to compensate for the lag time arising from the inertia of said paddle.