Pedestrian Detector

• US 20130136308A1
• Filed: 11/28/2011
• Published: 05/30/2013
• Est. Priority Date: 11/28/2011
• Status: Active Grant

• Current Assignee: National Chung-Shan Institute of Science and Technology
• Original Assignee: Chung-Shan Institute of Science & Technology Armaments Bureau MND (Government of The Republic of China)
• Inventors: Lo, Min-Fang, Fu, Li-Chen, Hsiao, Pei-Yung, Chan, Yi-Ming, Chuang, Li-An
• Granted Patent: US 8,649,564 B2
• Time in Patent Office: Days
• US Class Current: 382/104
• CPC Class Codes: G06V 10/50 by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads

Abstract

A pedestrian detector includes an image-taking mechanism and an image analyzer. The image-taking mechanism is provided on a vehicle and used to take an image in front of the vehicle. The image analyzer is connected to the image-taking mechanism and used to receive the image from the image-taking mechanism and analyze the image to determine whether the image conforms to a pedestrian to determine whether there is a pedestrian in front of the vehicle.

Claims

1. The pedestrian detector including:

an image-taking mechanism 1 provided on a vehicle 3 and used to take an image in front of the vehicle 3; and

an image analyzer 2 connected to the image-taking mechanism 1 and used to receive the image from the image-taking mechanism 1 and analyze the image to determine whether the image conforms to a pedestrian to determine whether there is a pedestrian in front of the vehicle 3.

2. The pedestrian detector according to claim 1, wherein the image-taking mechanism 1 includes a camera.

3. The pedestrian detector according to claim 1, wherein the image analyzer 2 includes:

a composite characteristic-extracting module 21 for extracting two types of characteristics of the image; and

a supported vector machine 22 for determining whether there is a pedestrian based on the two types of characteristics.

4. The pedestrian detector according to claim 3, wherein the composite characteristic-extracting module 21 includes:

a HOG characteristic-extracting module 211 for extracting silhouette characteristics of the image for later use;

a granule-based relative characteristic-extracting module 212 for extracting variations of brightness in granules 20 of the image; and

a characteristic-combination module 213 for combining the HOG characteristic-extracting module 211 with the granule-based relative characteristic-extracting module 212 to compare the two types of characteristics with each other.

5. The pedestrian detector according to claim 4, wherein the granule-based relative characteristic-extracting module 212 includes:

a granule brightness-averaging unit 2121 for averaging the brightness of each granule 20; and

a relative value-encoding unit 2122 for producing relative characteristics of the granules 20.

6. The pedestrian detector according to claim 4, wherein the granule-based relative characteristic-extracting module 212 extracts the variations of the brightness in the granules to produce a 10-dimensional vector to overcome errors related to the loss of the image.