Object identification device

1. An object identification device comprising:

a plurality of sensors, each of which:

detects an object, and

for each of a plurality of categories, determines and outputs a confidence indicating a degree of matching with a respective one from among the plurality of categories for the detected object; and

a processor that calculates an average confidence for each of the plurality of categories by weighted-averaging confidences output by the sensors, wherein the weighted-averaging comprises assigning different weights to at least two sensors from among the plurality of sensors for the same category from among the plurality of categories; and identifies a category from among the plurality of categories having a highest weighted-average confidence and categorizes the detected object into the identified category.

2. The object identification device according to claim 1 , wherein each of the plurality of sensors normalizes the confidence for each of the plurality of categories such that a summation of the confidence from the plurality of categories is equal to 1.

3. The object identification device according to claim 2 , wherein the plurality of sensors comprise a radar, which outputs the confidence to the processor, for each of the plurality of categories, and wherein the radar calculates the confidence from a reception level of an electric wave reflected on the object and received by the radar and a distance to the object.

4. The object identification device according to claim 2 , wherein the plurality of sensors comprise a camera, which outputs the confidence to the processor for each of the plurality of categories, and wherein the camera obtains the confidence based on a degree of matching with each of the plurality of categories, which is calculated by implementing a pattern matching between an image captured by the camera and the object in each of the plurality of categories.

5. The object identification device according to claim 1 , wherein the plurality of sensors comprise a radar, which outputs the confidence to the processor, for each of the plurality of categories, and wherein the radar calculates the confidence from a reception level of an electric wave reflected from the object and received by the radar and a distance to the object.

6. The object identification device according to claim 1 , wherein the plurality of sensors comprise a camera, which outputs the confidence to the processor for each of the plurality of categories, and wherein the camera obtains the confidence based on a degree of matching with each of the plurality of categories, which is calculated by implementing a pattern matching between an image captured by the camera and the object in each of the plurality of categories.

7. The object identification device according to claim 1 , wherein:

the plurality of categories comprises a pedestrian category, a motorcycle category, and a vehicle category, and

the processor normalizes the calculated average confidence for each of the plurality of categories such that a summation of the calculated average confidence for the plurality of categories is equal to 1 and in response to the object being in the pedestrian category, identify the object is a pedestrian ahead of a vehicle.

8. The object identification device according to claim 1 , wherein the processor applies pre-set weights to the calculated average confidence based on an authenticity of a respective sensor from among the plurality of sensors.

9. An object identification apparatus mounted in a vehicle comprising:

at least one camera which captures an image of an object in front of a vehicle and executes pattern matching of the captured image with respect to each of a plurality of categories to obtain a respective camera confidence value for each of the plurality of categories, and normalizes the obtained respective camera confidence value for each of the plurality of categories;

a radar which transmits a pulse to the object and receives an electric wave reflected by the object, calculates a respective radar confidence value for each of the plurality of categories, and normalizes the calculated respective radar confidence value for each of the plurality of categories; and

a processor, which weight-averages confidence for each of the plurality of categories using the respective camera confidence value output from the camera and the respective radar confidence value output from the radar and determines a category with a highest weight-average confidence as the category of the object from among the plurality of categories,

wherein the weight-average comprises assigning different weights to the respective radar confidence value from the radar and the respective camera confidence value from the camera with respect to at least one of the plurality of categories.

10. The object identification apparatus according to claim 9 , wherein the plurality of categories comprises a pedestrian category, a motorcycle category, and a vehicle category and wherein, in response to the object being determined to be in the pedestrian category, the processor identifies the object as a pedestrian ahead of the vehicle.

11. The object identification apparatus according to claim 10 , wherein different weights are set to each of the radar and the camera based on an authenticity of a respective device.

12. An object identification method comprising:

detecting an object by each of a plurality of sensors, and

determining and outputting by each of the plurality of sensors for each of a plurality of categories, a confidence indicating a degree of matching with a respective one from among the plurality of categories for the detected object; and

calculating, by a processor, an average confidence for each of the plurality of categories by weighted-averaging confidences output by the sensors, wherein the weighted-averaging comprises assigning different weights to at least two sensors from among the plurality of sensors in the same category, from among the plurality of categories;

identifying, by the processor a category having a highest weighted-average confidence; and

categorizing, by the processor, the detected object into the identified category.