Traffic sign recognition system and traffic sign recognition method

Abstract

A traffic signal recognition system method. The traffic signal recognition method includes: generating a first deformation value based on the difference between an image signal and at least one first pre-stored master feature; outputting a first standardized image signal and a first identification code based on the first deformation value corresponding to each of the at least one first pre-stored master feature; comparing the first standardized image signal with at least a portion of the plurality of first pre-stored traffic signals to generate at least one first matching value; and outputting a first traffic signal matching result and a second identification code based on the at least one first matching value. A first output element is used to confirm that when the first traffic signal matching result indicates that a first target traffic signal among the plurality of first pre-stored traffic signals and the first identification code are equal to the second identification code, outputting the first target traffic signal.

Claims

1. A traffic signal recognition system, comprising: a camera lens for capturing images of the surrounding environment to output an image signal; and a first storage element for storing a plurality of first pre-stored traffic signals and at least one first pre-stored main feature of the first pre-stored traffic signals; A first processor, connected to the camera lens and the first storage element, is configured to: generate a first deformation value based on the difference between the image signal and the at least one first pre-stored master feature; output a first standardized image signal and a first identification code based on the first deformation value corresponding to each of the at least one first pre-stored master feature, wherein the first standardized image signal is obtained by performing a first perspective transformation matrix operation on the image signal; compare the first standardized image signal with at least a portion of the first pre-stored traffic signals to generate at least one first matching value; and output a first traffic signal matching result and a second identification code based on the at least one first matching value; and a first output element, connected to the first processor, is configured to output the first target traffic signal when the first traffic signal matching result indicates that a first target traffic signal among the first pre-stored traffic signals and the first identification code is equal to the second identification code.

2. The traffic signal recognition system as claimed in claim 1 further comprises: a second storage element for storing a plurality of second pre-stored traffic signals and a second pre-stored main feature of one of the second pre-stored traffic signals; a second processor connected to the camera lens and the second storage element, configured to: generate a second deformation value based on the difference between the image signal and the second pre-stored main feature; output a second standardized image signal and a third identification code based on the second deformation value of the second pre-stored main feature, wherein the second standardized image signal is obtained by performing a second perspective transformation matrix operation on the image signal; compare the second standardized image signal with the second pre-stored traffic signals to generate a plurality of second matching values; and output a second traffic signal matching result and a fourth identification code based on the second matching values; and a second output element connected to the second processor, configured to output the second target traffic signal when the second traffic signal matching result indicates that a second target traffic signal among the second pre-stored traffic signals and the third identification code is equal to the fourth identification code.

3. The traffic signal recognition system as claimed in claim 2 further comprises: a third storage element for storing the at least one first pre-stored main feature and the second pre-stored main feature; a third processor connected to the camera lens and the third storage element, configured to: generate a third deformation value for each of the differences between the image signal and the at least one first pre-stored main feature and the second pre-stored main feature, and generate a main feature judgment result based on the third deformation value corresponding to each of the at least one first pre-stored main feature and the second pre-stored main feature; and a third output element connected to the third processor, the first output element and the second output element, wherein the third output element is configured to output the first target traffic signal when the main feature judgment result indicates that the image signal corresponds to the at least one first pre-stored main feature, and to output the second target traffic signal when the main feature judgment result indicates that the image signal corresponds to the second pre-stored main feature.

4. The traffic signal recognition system as claimed in claim 1, wherein the number of the at least one first pre-stored master features is multiple, and the first processor is configured to output the first standardized image signal and the first identification code when the first deformation value corresponding to a target master feature among the at least one first pre-stored master features is lower than a deformation tolerance value.

5. The traffic signal recognition system as claimed in claim 1, wherein the number of the at least one first matching value is multiple, and the first processor is configured to, when multiple of the at least one first matching values ​​are higher than a preset matching value, take the one of the first pre-stored traffic signals corresponding to the maximum matching value as the first traffic signal matching result.

6. A traffic signal recognition method, comprising: acquiring an image signal captured by a camera lens of the outside world by a first processor; generating a first deformation value for each of the image signal and at least one first pre-stored main feature of a plurality of first pre-stored traffic signals based on the difference between the image signal and the first deformation value corresponding to each of the at least one first pre-stored main feature; outputting a first standardized image signal and a first identification code by the first processor based on the first deformation value corresponding to each of the at least one first pre-stored main feature, wherein the first standardized image signal is obtained by performing a first perspective transformation matrix operation on the image signal; comparing the first standardized image signal with at least a portion of the first pre-stored traffic signals by the first processor to generate at least one first matching value; outputting a first traffic signal matching result and a second identification code by the first processor based on the at least one first matching value; and outputting the first target traffic signal when the first traffic signal matching result indicates that a first target traffic signal among the first pre-stored traffic signals and the first identification code is equal to the second identification code by a first output element.

7. The traffic signal recognition method as described in claim 6 further comprises: generating a second deformation value by a second processor based on the difference between the image signal and a second pre-stored main feature of one of a plurality of second pre-stored traffic signals; outputting a second standardized image signal and a third identification code by the second processor based on the second deformation value of the second pre-stored main feature, wherein the second standardized image signal is obtained by performing a second perspective transformation matrix operation on the image signal; comparing the second standardized image signal with the second pre-stored traffic signals by the second processor to generate a plurality of second matching values; outputting a second traffic signal matching result and a fourth identification code by the second processor based on the second matching values; and outputting the second target traffic signal by a second output element confirming that the second traffic signal matching result indicates that a second target traffic signal among the second pre-stored traffic signals and the third identification code is equal to the fourth identification code.

8. The traffic signal recognition method as described in claim 7 further comprises: generating a third deformation value for each of the image signal and the at least one first pre-stored main feature and the second pre-stored main feature using a third processor, and generating a main feature determination result based on the third deformation value corresponding to the at least one first pre-stored main feature and the second pre-stored main feature respectively; and outputting the first target traffic signal using a third output element when the main feature determination result indicates that the image signal corresponds to the at least one first pre-stored main feature, and outputting the second target traffic signal when the main feature determination result indicates that the image signal corresponds to the second pre-stored main feature.

9. The traffic signal recognition method as described in claim 6, wherein the number of the at least one first pre-stored master features is multiple, and the first processor outputs the first standardized image signal and the first identification code based on the first deformation value corresponding to each of the at least one first pre-stored master features, including: when the first deformation value corresponding to a target master feature among the at least one first pre-stored master features is lower than a deformation tolerance value, the first standardized image signal and the first identification code are output.

10. The traffic signal recognition method as described in claim 6, wherein the number of the at least one first matching value is multiple, and the first processor outputs the first traffic signal matching result and the second identification code based on the at least one first matching value, wherein when multiple of the at least one first matching values ​​are higher than a preset matching value, the one of the first pre-stored traffic signals corresponding to the maximum matching value is taken as the first traffic signal matching result.

Images