US5963663A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Land mark recognition method for mobile robot navigation
Assignee: Sony Corp
•
Filing: 1997-07-03
•
Granted: 1999-10-05
•
Expires: 2017-07-03
Abstract
Abstract A land mark recognition apparatus presenting an enhanced learning capability with its neural net. Color image data of a land mark picked up by a television camera is quantized in the space defined by hue and chroma. Specifically, the color image data defined by hue and chroma falls within an area A1, the data is quantized as red data. Similarly, color image data falls within areas A3 and A3, the data is quantized as green and blue data, respectively. The color image data, after being quantized into a small number of data, is stored in the neural net.
Inventors
- 👤 Jun Tani
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-06 02:39
CPC Classifications
G; G06; G06V; G06V10/00; G06V10/40
Claims Analysis
Total Claims: 11
Claim 1
Independent Claim
170 words
An object recognition apparatus for recognizing which of a plurality of possible reference objects is a particular reference object, comprising: imaging means for imaging the particular reference object to be recognized and outputting color image data corresponding to the particular reference object, quantizing means for quantizing individually the color image data of the particular reference object output by the imaging means, as data falling within one of segmented areas of a finite number in a predetermined color space, storing and learning means for learning and then storing, on an area-by-area basis, the individually quantized data provided by the quantizing means, comparison means for comparing, on an area-by-area basis, the area-by-area quantized data stored individually in the storing and learning means with the area-by-area quantized data provided by the quantizing means, and determining means for exclusively determining which of the plurality of possible reference objects is the particular reference object based on the collective comparison result for the areas provided by the comparison means that makes comparisons on an area-by-area basis.
Claim 2
Dependent Claim
170 words
An object recognition apparatus for recognizing which of a plurality of possible reference objects is a particular reference object, comprising: imaging means for imaging the particular reference object to be recognized and outputting color image data corresponding to the particular reference object, quantizing means for quantizing individually the color image data of the particular reference object output by the imaging means, as data falling within one of segmented areas of a finite number in a predetermined color space, storing and learning means for learning and then storing, on an area-by-area basis, the individually quantized data provided by the quantizing means, comparison means for comparing, on an area-by-area basis, the area-by-area quantized data stored individually in the storing and learning means with the area-by-area quantized data provided by the quantizing means, and determining means for exclusively determining which of the plurality of possible reference objects is the particular reference object based on the collective comparison result for the areas provided by the comparison means that makes comparisons on an area-by-area basis.
Claim 3
Dependent Claim
30 words
An object recognition apparatus according to claim 1, wherein the color space is defined by hue and chroma from among the three attributes of color, namely, hue, chroma and brightness.
Claim 4
Dependent Claim
85 words
An object recognition apparatus according to claim 1, wherein the storage means stores, on an area-by-area basis, the quantized data corresponding to the reference object to be recognized, as a connectivity weight between neurons in a neural net, the comparison means compares, on an area-by-area basis, the quantized data in terms of the fire condition of each neuron in the neural net, and an associative memory net that recognizes a reference object currently being imaged is constituted by the storage means and the comparison means.
Claim 5
Dependent Claim
18 words
An object recognition apparatus according to claim 3, wherein the associative memory net comprises a Hopfield neural net.
Claim 6
Dependent Claim
28 words
An object recognition apparatus according to claim 1, wherein the determining means comprises a winner-take-all type neural network for converging the comparison result provided by the comparison means.
Claim 7
Dependent Claim
69 words
An object recognition apparatus according to claim 1 further comprising a recurrent neural network which has a context loop which internally reflects history to learn the time-series order of appearance of the particular reference object determined by the determining means and to predict another particular reference object that is one of the plurality of possible reference objects to subsequently appear based on the determination result of the determining means.
Claim 8
Dependent Claim
37 words
An object recognition apparatus according to claim 1, wherein the particular reference object is used as a land mark by which a mobile robot moving autonomously along a path determines the current position of the mobile robot.
Claim 9
Dependent Claim
128 words
An object recognition method for recognizing which of a plurality of possible reference objects is a particular reference object, comprising the steps of: quantizing means for quantizing individually the color image data which is obtained by imaging, the particular reference object to be recognized, as data falling within one of segmented areas of a finite number in a predetermined color space, learning and then storing the individually quantized data provided by the quantizing means, on an area-by-area basis, comparing, on an area-by-area basis, the area-by-area quantized data stored individually with the area-by-area quantized data obtained through imaging, and exclusively determining which of the plurality of possible reference objects is the particular reference object based on the collective comparison result for the areas resulting from the area-by-area basis comparison.
Claim 10
Dependent Claim
128 words
An object recognition method for recognizing which of a plurality of possible reference objects is a particular reference object, comprising the steps of: quantizing means for quantizing individually the color image data which is obtained by imaging, the particular reference object to be recognized, as data falling within one of segmented areas of a finite number in a predetermined color space, learning and then storing the individually quantized data provided by the quantizing means, on an area-by-area basis, comparing, on an area-by-area basis, the area-by-area quantized data stored individually with the area-by-area quantized data obtained through imaging, and exclusively determining which of the plurality of possible reference objects is the particular reference object based on the collective comparison result for the areas resulting from the area-by-area basis comparison.
Claim 11
Dependent Claim
58 words
An object recognition method according to claim 8 further comprising the step of learning the time-series order of appearance of the particular reference object exclusively determined, based on the comparison result through the area-by-area comparison, and predicting another reference object that is one of the plurality of possible reference objects to subsequently appear, based on the determination result.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2002-10-04 | 2003-04-05 | 2003-10-05 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2006-10-04 | 2007-04-05 | 2007-10-05 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2010-10-04 | 2011-04-05 | 2011-10-05 | $7700.00 | Window Closed / Lapsed |