Pulse-coupled automatic object recognition system dedicatory clause

• US 5,664,065 A
• Filed: 06/17/1996
• Issued: 09/02/1997
• Est. Priority Date: 06/17/1996
• Status: Expired due to Fees

• Current Assignee: United States Department of the Army
• Original Assignee: The United States Of America As Represented By The Secretary Of The Army
• Inventors: Johnson, John L.
• Primary Examiner(s): Hafiz, Tariq R.
• Assistant Examiner(s): CHUNG, JI YONG DAVID
• Application Number:
  US08/668,029
• Time in Patent Office: 442 Days
• Field of Search: 382/158, 382/159, 359/558, 359/576, 359/559, 395/22, 395/24, 395/21, 395/27, 395/20
• US Class Current: 706/20
• CPC Class Codes: G06N 3/0675 using electro-optical, acousto-optical or opto-electronic means
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 10/895 the filter being related to phase processing, e.g. phase-only filters
• CPC Class Codes: G06V 30/194 References adjustable by an adaptive method, e.g. learning

Abstract

The pulse-coupled automatic recognition system processes information utilizing parallel neural network and transmits information to and from the processing sites as time signals which encode in their pulse phase structure the geometrical content of the spatial distributions of light. A locating device then measures the coordinates of the spatial distributions (pulsating segment) and displays the coordinates of the pulsating segment on a screen for observation. The pulsating segment indicates the presence and location of the component of the input scene that corresponds to the segment.

Claims

1. A pulse-coupled automatic object recognition system for detecting and recognizing pre-selected objects in a scene, said system comprising:

a first and a second pulse-coupled neural networks (PCNN'"'"'s), said PCNN'"'"'s being capable of producing corresponding electronic pulse patterns in response to optical inputs thereto, said optical inputs being descriptive of the scene;

a means for producing reference time signals, said time signals being characteristic of the portions of the scene that correspond to said time signals;

a means for detecting light;

a gating means simultaneously coupled between said second PCNN, said light-detecting means and said time signal-producing means for selectively allowing, in response to time signal inputs from said signal-producing means, the transmission of light from said second PCNN onto said light-detecting means, said transmitted light being indicative of the pre-selected object and a measuring means coupled to said light-detecting means to measure the coordinates of light distribution on said light-detecting means, thereby yielding the positional information of the pre-selected object in the input scene.

2. A pulse-coupled automatic object recognition system as set forth in claim 1, wherein said first and second PCNN'"'"'s have the identical number and geometrical arrangement of pixels.

3. A pulse-coupled automatic object recognition system as set forth in claim 2, wherein said system further comprises a means to illuminate the pulse patterns on said first and second PCNN'"'"'s, thereby producing optical images of said pulse patterns.

4. An automatic object recognition system as set forth in claim 3, wherein said means for producing reference time signals comprises a single-pixel detector capable of generating time signals in response to optical inputs thereto;

a first focussing lens positioned to receive the optical image of the electronic pulse patterns from said first PCNN and focus the image onto said single-pixel detector; and

a time correlator, said correlator being coupled to said single-pixel detector to receive therefrom and store the time signals, said correlator further transmitting pre-selected time signals.

5. An automatic object recognition system as set forth in claim 4, wherein said light-detecting means is a charge-coupled device (CCD) array.

6. An automatic object recognition system as set forth in claim 5, wherein said gating means comprises a second focussing lens positioned to receive from said second PCNN the optical image of the electronic pulse patterns;

a collimating lens for collimating incident light and transmitting the light to impinge on said CCD array; and

a time gate, said gate being coupled to said time correlator and being further positioned between said second focussing lens and collimating lens selectively to open, in response to time signal inputs from said time correlator, thereby selectively allowing the passage of focussed light from said second focussing lens onto said collimating lens.

7. An object recognition system as set forth in claim 6, wherein said measuring means is a centroid locator.

8. An object recognition system as set forth in claim 7, wherein said system still further comprises a means for displaying, said displaying means being coupled to said centroid locator for receiving from said locator the positional information of the pre-selected object and displaying the information for visual observation.

9. An object recognition system as set forth in claim 8, wherein said first, second PCNN'"'"'s and said CCD array all have the identical number and geometrical arrangement of pixels.

10. A pulse-coupled automatic object recognition system for locating and recognizing pre-selected objects in a scene, said system comprising:

a first and a second pulse-coupled neural networks (PCNN'"'"'s), said PCNN'"'"'s being capable of generating corresponding electronic pulse patterns in response to optical inputs thereto, said optical inputs being descriptive of the scene and said pulse patterns modulating any beam of light passing therethrough;

a single-pixel detector for producing time signals in response to input images focussed thereonto;

a first focussing lens positioned between said first PCNN and said detector to receive the modulated light emanating from said first PCNN and focus said modulated light to be incident on said detector;

a time correlator coupled to said single-pixel detector for receiving therefrom the time signals, said time correlator being capable of further transmitting pre-selected time signals that correspond to pre-selected objects in the scene;

a second focussing lens positioned to receive and focus the modulated light emanating from said second PCNN;

a collimating lens;

a time gate located between said second focussing lens and collimating lens to receive the focussed modulated light from said second focussing lens and transmit the focussed light to be incident on said collimating lens, said lime gate being further coupled to said time correlator to receive therefrom pre-selected time signals and opening selectively in response to the pre-selected time signals;

a detector array for detecting the collimated light incident thereon from said collimating lens; and

a centroid locator coupled to measure the coordinates of the light distribution on said detector array to provide the locational information of the pre-selected objects in the scene.

11. A pulse-coupled automatic object recognition system as set forth in claim 10, wherein each of said first and second PCNN'"'"'s is illuminated by its own light source.

12. An automatic object recognition system as set forth in claim 11, wherein said system further comprises a screen connected to said centroid locator for receiving the object locational information from said locator and displaying the position of the object on said screen.

13. An automatic object recognition system as set forth in claim 12, wherein said display of the pre-selected object image has equivalent position with respect to the entire screen as does the object with respect to the entire input scene.

14. An object recognition system as set forth in claim 13, wherein said first, second PCNN'"'"'s and said detector array all have the identical number and geometrical arrangement of pixels.

15. An electro-optical pulse-coupled neural network (PCNN) having a plurality of pixels, said PCNN comprising:

a first spatial light modulator (SLM) for receiving and modulating input light passing therethrough;

a first detector array for detecting light incident thereon and producing electronic video signals in response to the light;

a first re-imaging lens positioned linearly between said first SLM and said first detector array for receiving the modulated light from said first SLM and focussing the light to be incident on said first detector array;

a second detector array;

a threshold discriminator coupled between said first detector array and said second detector array to produce a causal signal in response to the inputs from said first and second detector arrays;

a pulse former coupled to said discriminator to generate at least one output pulse in response to said causal signal;

a threshold means coupled to receive said output pulse and produce therefrom a threshold signal, said threshold means comprising a first phospher-coated glass plate;

a second SLM coupled to receive said output pulse from said pulse former;

a first illuminator positioned to illuminate said second SLM;

a second re-imaging lens located to focus the image of said output pulse from said second SLM onto said first phospher-coated glass plate, said first plate exhibiting phospher decay in response to the focussed image of said output pulse; and

a third re-imaging lens positioned to focus the decaying image onto said second detector array, said second detector array producing the threshold signal in response to said decaying image, said threshold means further transmitting said threshold signal to said discriminator; and

a means for summing the output pulses of several of the pixels to produce a linking input suitable for inputting to said first SLM.

16. An electro-optical PCNN as described in claim 15, wherein said summing means comprises a third SLM coupled to receive said output pulse from said pulse former;

a second illuminator suitable for illuminating said third SLM;

a second phospher-coated glass plate capable of undergoing phospher decay in response to image focussed thereon;

a first summing lens coupled between said third SLM and said second glass plate to perform summation of the output pulses of a pre-selected set of several pixels of said third SLM and focus the sum onto said second glass plate to cause the phospher decay thereon;

a second summing lens;

a bias positioned between said second glass plate and said second summing lens to add a pre-determined bias to said decaying image, said second summing lens focussing said biased image onto said first SLM.

17. An electro-optical PCNN as described in claim 16, wherein said summing means further comprises an intensity mask, said mask movably positioned between said third SLM and said second glass plate to manipulate the size of said pre-selected set of several pixels.

18. An electro-optical PCNN as described in claim 17, wherein said bias comprises a beamsplitter located between said second glass plate and said second summing lens and a source of constant-intensity light, said source being suitably positioned to illuminate said beamsplitter.