TWI469064B - Video recognizing system and method - Google Patents

Description

Image recognition system and method

The present invention relates to an image recognition system and method.

The conventional digital image monitoring system captures only a single camera for the same scene, and recognizes the image captured by the single camera. In order to carry out the identification more accurately, the resolution of the camera generally has high requirements. However, since the resolution of the image obtained by the camera with high resolution is also high, the image file returned by the camera is large, and at this time, it is necessary to compress the high-resolution image and then return it. Pass to the backend processing system. This will result in a longer processing time.

In view of the above, it is necessary to provide a faster image recognition system and method.

An image recognition system includes: a first camera for photographing a scene to obtain a first image; and a second camera for photographing the scene to obtain a second image The resolution of the second image is higher than the resolution of the first image, the first image has the same content as the second image, and the size of the first image is smaller than the size of the second image; a detecting unit, configured to detect the first image to obtain a key in the first image An image of the region; a landmark determining unit configured to obtain a coordinate range of the key region in the first image, and obtain a key region in the second image according to a size ratio between the second image and the first image a coordinate range; a capture unit configured to extract an image of the corresponding region from the second image according to a coordinate range of the key region in the second image, and according to the first image and the second image The size ratio is correspondingly reduced, the image of the reduced key area is used to replace the key area in the first image to obtain a third image, and an identification unit is configured to identify the third image.

An image recognition method includes: a first camera and a second camera simultaneously shooting the same scene to obtain a first image and a second image, wherein the first image has a comparison with the second image Low resolution, the first image has the same content as the second image and the size of the first image is smaller than the size of the second image; the first image is transmitted to an image processing system; the image processing The system detects the first image to obtain an image of a key area in the first image; the image processing system correspondingly obtains a coordinate range corresponding to the key area in the first image, and according to the second figure Corresponding to the size ratio between the image and the first image, the coordinate range of the key region in the second image is obtained; the image processing system correspondingly extracts the image of the key region from the second image according to the obtained coordinate range, and It is based on the ratio of the size between the first image and the second image Should be reduced, the image of the reduced key area is used to replace the image of the key area in the first image to obtain the third image; and the image processing system recognizes the third image.

The image recognition system detects the first image with a lower resolution to obtain a coordinate range of the key region, and obtains a coordinate range of the corresponding region in the second image according to the obtained coordinate range. Then, the area of the corresponding coordinate range in the second image is captured and reduced accordingly. The image of the reduced area is used to replace the image of the key area in the first image. Since the second image has a higher resolution, the subsequent identification process can be more accurate, and at the same time, only the corresponding region in the second image is captured, and the entire second image is prevented from being transmitted to the image processing. The processing time caused by the system is too long.

10‧‧‧First camera

20‧‧‧Second camera

30‧‧‧Image Processing System

301‧‧‧Detection unit

302‧‧‧ coordinate determination unit

303‧‧‧Capture unit

304‧‧‧ Identification unit

50‧‧‧Photographed scene

51‧‧‧ first image

52‧‧‧ second image

1 is a block diagram of a preferred embodiment of an image recognition system of the present invention.

2 is a block diagram of the image processing system of FIG. 1.

Figure 3 is a schematic illustration of a scene taken.

4A and 4B are schematic views of an image obtained by photographing a scene to be taken by the first and second cameras.

FIG. 5 is a schematic diagram of the detection unit of FIG. 2 detecting an image obtained by the first camera in FIG. 4. FIG.

6 is a flow chart of a preferred embodiment of the image recognition method of the present invention.

Referring to FIG. 1, a preferred embodiment of the image recognition system of the present invention includes a first shot. The camera 10, a second camera 20, and an image processing system 30. The first camera 10 has a lower resolution than the second camera 20, and the captured image has a lower resolution than the camera 20. In the present embodiment, the first camera 10 and the second camera 20 are configured to capture the same scene at the same time to obtain two images having the same content but different resolutions. The resolution of the first image captured by the first camera 10 is lower than the resolution of the second image captured by the second camera 20, and the size of the first image is also smaller than the second image. The size of the image.

As shown in FIG. 2 , the image processing system 30 includes a detecting unit 301 , a label determining unit 302 , a capturing unit 303 , and an identifying unit 304 .

The first camera 10 transmits the captured first image to the image processing system 30.

The detecting unit 301 is configured to detect the first image to obtain an image of a key area in the first image. In this embodiment, the detecting unit 301 is a face detecting unit that performs face detection on the first image to obtain an image of the face region. In other embodiments, the detecting unit 301 can also be a license plate detecting unit. At this time, the detecting unit 301 obtains an image of the license plate area.

The coordinate determining unit 302 is configured to process the image of the face region detected by the detecting unit 301 to obtain a coordinate range of the face region in the first image, and correspondingly obtain the second camera 20 The coordinate range of the face region in the second image stored in . In the present embodiment, it is assumed that the coordinates of the center point of the first image and the second image are (0, 0). Since the content of the first image and the second image are the same, the coordinate determining unit 302 can obtain the coordinate range of the face region portion in the second image according to the size ratio between the first image and the second image. For example, if the first image and the second image The ratio between the dimensions of the image is 1:3, and the relationship between the coordinates of the face region in the second image and the coordinates of the face region in the first image is 3:1.

The capturing unit 303 is configured to capture a corresponding area in the second image according to the coordinate range, and replace the captured area in the second image with the face area in the first image, and then replace the first image. The face area to get a third image. The face region of the third image obtained at this time has a higher resolution, while the other regions have a lower resolution.

The identification unit 304 is configured to identify a facial feature in the third image, such as comparing with a pre-stored image, to determine whether the person photographed by the first camera 10 and the second camera 20 is a legitimate user. . Among them, the identification technology is a well-known technique in the art, for example, by comparing the primitives of the same area between the third image and the pre-stored image, the similarity between the two images can be known, when the similarity is When a certain predetermined value is exceeded, the person in the third image is considered to be the same as the person in the pre-stored image.

The image recognition system detects the first image having a lower resolution to obtain a coordinate range of the face region, and correspondingly obtains a coordinate range of the face region in the second image. Then, corresponding regions in the second image are extracted according to the obtained coordinate range to synthesize the third image. The face region in the third image has a higher resolution, so that the subsequent identification process can be more accurate, and at the same time, only the corresponding region in the second image is captured, and the entire second image is avoided. The processing time caused by the image processing system is too long.

The specific working principle of the image recognition system of the present invention will be described below by way of an example.

Referring to FIG. 3 and FIG. 4A-4B together, the first camera 10 and the second camera 20 simultaneously capture a scene 50 to obtain the first image 51 and the second image 52. its The first image 51 and the second image 52 have the same content. However, the first image 51 has a lower resolution and size than the second image 52. The first image 51 is transmitted to the image processing system 30 at the back end.

Referring to FIG. 5 together, the detecting unit 301 detects the first image 51, and obtains three face regions in the first image 51, as shown by the dashed box in FIG. The coordinate determining unit 302 correspondingly obtains the coordinate range corresponding to the three face regions. Thereafter, the coordinate determining unit 302 obtains the coordinate range of the three face regions in the second image 52 according to the size ratio between the first image 51 and the second image 52.

The capturing unit 303 extracts three face regions from the second image 52 stored in the second camera 20 according to the coordinate range obtained by the coordinate determining unit 302, and extracts the three face regions of the captured The image corresponds to zoom out. The image of the reduced three face regions is used to replace the three face regions in the first image 51 to obtain the third image. At this time, the face area in the third image has a higher resolution, and the other areas have a lower resolution. The identification unit 304 identifies the third image.

Referring to FIG. 6, a preferred embodiment of the image recognition method of the present invention includes the following steps:

Step S1: The first camera 10 and the second camera 20 capture the same scene 50 at the same time to obtain the first image 51 and the second image 52, respectively. The first image 51 and the second image 52 have the same content. However, the first image 51 has a lower resolution and a smaller size than the second image 52.

Step S2: The first image 51 is transmitted to the image processing system 30 at the back end.

Step S3: the detecting unit 301 detects the first image 51, and obtains the first There are three face regions in the image 51.

Step S4: the coordinate determining unit 302 correspondingly obtains the coordinate range corresponding to the three face regions, and obtains three of the second images 52 according to the size ratio between the first image 51 and the second image 52. The range of coordinates of the face area.

Step S5: The capturing unit 303 extracts three facial images from the second image 52 stored in the second camera 20 according to the coordinate range obtained by the coordinate determining unit 302, and obtains three facial images. The image is correspondingly reduced in accordance with the size ratio between the first image 51 and the second image 52. The reduced three face images are images for replacing the coordinate ranges corresponding to the three face regions in the first image 51 to obtain the third image. At this time, the face area in the third image has a higher resolution, and the other areas have a lower resolution.

Step S6: The identification unit 304 identifies the third image.

In summary, the present invention has indeed met the requirements of the invention patent, and has filed a patent application according to law. The above is only the preferred embodiment of the present invention, and the scope of the present invention is not limited to the above-described embodiments, and equivalent modifications or variations made by those skilled in the art in accordance with the spirit of the present invention are It should be covered by the following patent application.

10‧‧‧First camera

20‧‧‧Second camera

30‧‧‧Image Processing System

Claims (6)

An image recognition system includes: a first camera for photographing a scene to obtain a first image; and a second camera for photographing the scene to obtain a second image The resolution of the second image is higher than the resolution of the first image, the first image has the same content as the second image, and the size of the first image is smaller than the size of the second image; a detecting unit, configured to detect the first image to obtain an image of a key area in the first image; and a target determining unit, configured to obtain a coordinate range of the key area in the first image, and according to a size ratio between the second image and the first image corresponds to a coordinate range in which the key region in the second image is located; a capture unit configured to extract the second image from the coordinate range of the key region in the second image Extracting an image of the corresponding area and reducing it according to a size ratio between the first image and the second image, the image of the reduced key area being used to replace the key area in the first image Obtaining a third image; and identifying Element, for identifying the third image. The image recognition system of claim 1, wherein the detecting unit is a face detecting unit, and the key area is a face area. The image recognition system of claim 1, wherein the detecting unit is a license plate detecting unit, and the key area is a license plate area. An image recognition method includes: a first camera and a second camera simultaneously shooting the same scene to obtain a first image and a second image, wherein the first image has a comparison with the second image Low resolution Degree, the first image has the same content as the second image and the size of the first image is smaller than the size of the second image; the first image is transmitted to an image processing system; the image processing system The first image is detected to obtain an image of a key area in the first image; the image processing system correspondingly obtains a coordinate range corresponding to the key area in the first image, and according to the second image and the The size ratio between the images corresponds to the coordinate range in which the key region in the second image is located; the image processing system correspondingly extracts the image of the key region from the second image according to the obtained coordinate range, and according to the The size ratio between the image and the second image is correspondingly reduced, and the image of the reduced key area is used to replace the image of the key area in the first image to obtain the third image; The image processing system recognizes the third image. The image recognition method of claim 4, wherein the key area is a face area. The image recognition method of claim 4, wherein the key area is a license plate area.