Three-dimensional object recognizing system

1. A three-dimensional object recognizing system comprising:

a distance image generating portion for generating a distance image including three-dimensional distance data of a picked-up object, by using images picked up by an imaging portion; and

a computer coupled to said distance image generating portion, comprising:

a grouping processing portion for grouping the distance data indicating the picked-up object on the distance image;

an input value setting portion for setting an area having a previously set predetermined profile and containing distance data group for the picked-up object on the distance image, and setting input values having an are distance data for every one of a plurality of partitioned areas of the distance image;

a computing portion for computing output values having a pattern that responds to a previously set three-dimensional object by using a neural network that has at least the input values as inputs to an input layer; and

a discriminating portion for discriminating the three-dimensional object based on the pattern of the output values computed by the computing portion.

2. A three-dimensional object recognizing system according to claim 1 , wherein the input value setting portion sets a rectangular profile surrounding the distance data group on the distance image, and sets the plurality of partitioned areas, each side of which circumscribes any pixel of the distance data group, by reducing the rectangular profile on a line by line basis.

3. A three-dimensional object recognizing system according to claim 1 , wherein the input value setting portion sets a mean value of distance data for every one of the plurality of areas as an element of the input values.

4. A three-dimensional object recognizing system according to claim 3 , wherein the input value setting portion calculates the mean value of distance data for every one of the plurality of areas at a subpixel level.

5. A three-dimensional object recognizing system according to claim 1 , wherein at least one weighting factor set in the neural network used by the computing portion are set by a genetic algorithm that employs a plurality of previously prepared distance images as a learning image.

6. A three-dimensional object recognizing system according to claim 5 , further comprising a control parameter learning unit configured to set the at least one weighting factor based on the previously set three-dimensional object.

7. A three-dimensional object recognizing system according to claim 6 , wherein the control parameter learning unit includes an evolution computing portion configured to adjust the weighting factors based on the genetic algorithm.

8. A three-dimensional object recognizing system according to claim 1 , wherein the pattern corresponding to the three-dimensional object is set as a different pattern for every type of the three-dimensional object, and the discriminating portion specifies the type of the three-dimensional object based on the pattern.

9. A three-dimensional object recognizing system according to claim 1 , wherein the neural network comprises a hierarchical type neural network comprising the input layer, a middle layer, and an output layer, and

wherein the output values are output from the output layer.

10. A three-dimensional object recognizing system according to claim 9 , wherein each of the input layer, the middle layer, and the output layer includes a plurality of nodes.

11. A three-dimensional object recognizing system according to claim 10 , wherein the neural network is configured to apply a first weighting factor between nodes of the input layer and the middle layer,

wherein the neural network is configured to apply a second weighting factor between nodes of the middle layer and the output layer, and

wherein the first and second weighting factors are based on the previously set three-dimensional object.

12. A three-dimensional object recognizing system according to claim 11 , wherein the first and second weighting factors set in the neural network used by the computing portion are set by a genetic algorithm that employs a plurality of previously prepared distance images as a learning image.

13. A three-dimensional object recognizing system according to claim 11 , further comprising a control parameter learning unit configured to set the first and second weighting factors based on the previously set three-dimensional object.

14. A three-dimensional object recognizing system according to claim 13 , wherein the control parameter learning unit includes an evolution computing portion configured to adjust the first and second weighting factors based on the genetic algorithm.

15. A three-dimensional object recognizing system according to claim 1 , wherein the grouping processing portion is adapted to calculate a planar expression as a ground surface and to extract the distance data which are located lower than the planar expression.

16. A three-dimensional object recognizing system according to claim 1 , wherein a range including the group comprises a minimum rectangle.

17. A three-dimensional object recognizing method comprising:

generating in a distance image generating portion a distance image, including three-dimensional distance data of a picked-up object, by using images picked up by an imaging portion; and

in a computer coupled to said distance image generating portion:

grouping the distance data indicating the picked-up object on the distance image;

setting an area having a previously set predetermined profile and containing a distance data group for the picked-up object on the distance image, and setting input values having an area distance data for every one of a plurality of partitioned areas of the distance image;

computing output values having a pattern that responds to a previously set three-dimensional object by using a neural network comprising an input layer and including at least the input values as inputs to the input layer; and

discriminating the three-dimensional object based on the pattern of the output values computed by the computing portion.

18. A three-dimensional object recognizing method according to claim 17 , wherein at least one weighting factor is set in the neural network used by the computing portion are set by a genetic algorithm that employs a plurality of previously prepared distance images as a learning image.

19. A three-dimensional object recognizing method according to claim 18 , wherein the at least one weighting factor is based on the previously set three-dimensional object.

20. A three-dimensional object recognizing method according to claim 18 , further comprising adjusting the at least one weighting factor based on the genetic algorithm.