Method and system for detecting object(s) adjacent to vehicle
AssigneeHua-Chuang Automobile Information Technical Center Co., Ltd.
InventorsYen-Lin Chen, Chung-Yuan Chang, Kuan-Kai Liao, Chao-Wei Yu, Ming Chen
Life of this patent
Estimating the term…
Abstract
A method for detecting one or more objects adjacent to a vehicle includes capturing a image of an object adjacent to the vehicle; determining a driving area in the image; cutting the driving area to form an identification window; selecting an identification area in the identification window; accessing a plurality of object image data in a memory to compare the plurality of object image data with the identification area; and identifying a specific object in the identification area by the processor.
Claims8
A method for detecting one or more objects adjacent to a vehicle, comprising:
- capturing, by an image capturing device, an image of an object adjacent to the vehicle;
- determining, by a processor, a driving area in the image, wherein a horizontal line of the image is a middle line of the driving area, and a height is extending upwardly and downwardly from the middle line along a direction perpendicular to the middle line;
- cutting, by the processor, the driving area to form an identification window;
- grayscaling, by the processor, the identification window to obtain a grayscaled identification window;
- applying, by the processor, gamma correction to the grayscaled identification window to obtain a corrected grayscaled identification window;
- selecting, by the processor, an identification area in the corrected grayscaled identification window;
- adjusting, by the processor, a range of the identification area by image pyramid techniques;
- accessing, by the processor, a plurality of object image data in a memory to compare the plurality of object image data with the identification area;
- and
- identifying, by the processor, a specific object in the identification area with an image processing method using histogram of oriented gradients in response to comparing the plurality of object image data with the identification area, wherein gradient vector generated by the histogram of oriented gradients are normalize or weighted to emphasize main edge vector features in consecutive cells.
A system for detecting one or more objects adjacent to a vehicle, installed in the vehicle, the system comprising:
- an image capturing device configured to capture an image of an object adjacent to the vehicle;
- a memory configured to store a plurality of object image data;
- and
- a processor electrically connected to the image capturing device, communicationally connected to the memory, and configured to:
- determine a driving area in the image, wherein a horizontal line of the image is a middle line of the driving area, and a height is extending upwardly and downwardly from the middle line along a direction perpendicular to the middle line;
- cut the driving area to form an identification window;
- grayscale the identification window to obtain a grayscaled identification window;
- apply gamma correction to the grayscaled identification window to obtain a corrected a corrected grayscaled identification window;
- select an identification area in the corrected grayscaled identification window;
- adjust a range of the identification area by image pyramid techniques;
- access the plurality of object image data in the memory to compare the plurality of object image data with the identification area;
- and
- identify a specific object in the identification area with an image processing method using histogram of oriented gradients in response to comparing the plurality of object image data with the identification area, wherein gradient vector generated by the histogram of oriented gradients are normalize or weighted to emphasize main edge vector features in consecutive cells.
Dependencies are inferred from the claim text.