CN102360424A - Glare prevention method based on shadow image cutting algorithm - Google Patents

Abstract

The invention discloses an anti-glare method based on a shadow image cutout algorithm. A CCD video acquisition module installed on the projector body is used to obtain the image of the speaker in front of the projector and its shadow on the screen. By distinguishing the projector The image of the speaker in front and the shadow position projected on the screen, and then estimate the relative position of the speaker to realize the shadow image removal algorithm. Image processing and data communication are convenient, easy to install, and realize the technical effect of being ready to use. When the position of the speaker changes, the position of the shadow on the screen also changes, and the edge position of the two changes immediately, thus triggering a new positioning operation, and there is no disadvantage of staying in multiple shadows. The algorithm is concise, has the advantages of simple operation, low hardware system requirements, low cost, wide application, etc., and has good economic benefits and market prospects.

Description

An anti-glare method based on shadow image matting algorithm

technical field

The present invention relates to an anti-glare method for a projector realized by a shadow image cutout algorithm, in particular to a method for real-time acquisition and processing by a CCD module to distinguish the image of a lecturer in front of the projector and the shadow position projected on the screen, and then The invention relates to an anti-glare method for estimating the relative position of a lecturer and controlling a projector not to irradiate the lecturer, especially eyes, and belongs to the technical field of projector control.

Background technique

Projectors are widely used in various conferences and teaching occasions, but the strong light of the projector strongly stimulates the speaker. After the speaker walks out of the projection area, the normal vision cannot be restored in a short time, which brings inconvenience to the speech. This is the glare problem of projectors. At present, none of the existing projector products provides a mature commercial solution for anti-glare.

In order to solve the problem of projector glare, there are currently several technical solutions. The Chinese invention patent "Intelligent Video Projection System for Protecting Speaker's Eyes" (Publication No.: CN101872329A) discloses an anti-glare method that collects surveillance video and detects the position of the speaker's eyes.

Another technical solution is a Chinese invention patent "device with projection control function and control method, computer, projector" (authorized announcement number: CN 101452193B) which discloses an actual environment image of the projected area of the projector through affine processing , and compare it with the original image to be projected to determine the position of the speaker, so as to eliminate the impact of projected light on the eyes of the speaker during the projection process after anti-glare treatment.

The above two methods are the two ideas that can be retrieved at present, but both have defects.

For the human eye detection method, before the position of the human eye is detected, the projector projects images with target text and patterns, and these images will be directly superimposed on the face of the speaker. As we all know, computer image processing algorithms are currently used to detect, The technology of face tracking has very high requirements on face images. Generally, a certain accuracy rate can be met by comparing information such as face skin color and features. After superimposing the target text and pattern information on the projector, the face detection and tracking is more difficult.

For the affine comparison method, first of all, it is necessary to ensure that the image acquisition module is installed very close to the projector screen, otherwise the collected image will not only contain the shadow image of the speaker, but also the image of the speaker himself. In the actual environment, it is very difficult to compare the original images to be projected, and the results obtained by the comparison cannot be directly used for anti-glare treatment.

It can be seen that the above two technical solutions are not only difficult to process, but also have high computational complexity and unstable performance. The present invention discloses a new type of method for distinguishing between the image of the speaker in front of the projector and the position of the shadow image projected on the screen. , and then estimate the relative position of the speaker, and control the anti-glare method that the projector does not illuminate the speaker, especially the eyes, which has the advantages of good realizability, simple processing, and low algorithm complexity, and can be widely used in projector production manufacturing field.

Contents of the invention

The purpose of the present invention is to provide an anti-glare method for a projector. When a lecturer enters the irradiation area of the projector when giving instructions in front of the projector or in front of the curtain, he is easily irradiated by the high-intensity light of the projector and damages the eyes. The invention realizes the anti-glare function by modifying the shadow part of the lecturer's occluded content to black and reducing the light intensity of the projector irradiating the human eyes. The technical difficulty of anti-glare lies in: how to identify the position of the human eye through the image recognition method, and how to dynamically track the position of the human eye.

In order to achieve the above technical goals, the present invention uses a CCD video acquisition module installed on the projector body to acquire the image of the speaker in front of the projector and its shadow on the screen.

By distinguishing the image of the speaker in front of the projector and the shadow position projected on the screen, the relative position of the speaker is estimated to realize the shadow image removal algorithm.

In fact, the image content obtained by the CCD video acquisition module is very complex, including the speaker image, shadows, the entire projected content area, and the non-content area outside the screen. Therefore, it is necessary to distinguish the above image contents.

The flow process of the technical solution adopted in the present invention is as follows:

Step 1. Image recognition process

(1) Get the background image

Since the projected content area of the whole field generally changes less, and the non-content area outside the curtain also changes less, it can be regarded as a background image, and a smooth and filtered background image can be obtained by repeatedly superimposing multiple frames of video images;

(2) Divide to obtain the projected content area of the audience

Due to the high brightness of the projected content area of the whole field, and the low brightness of the non-content area outside the screen, the rectangular area of the projected content area of the whole field can be segmented by calculating the brightness gradient map of the background image;

(3) Identify shadows

Under normal circumstances, the shadow on the screen is relatively easy to detect, and the shadow area can be directly distinguished by setting the brightness threshold;

(4) Detect the speaker image area

Through the difference operation with the background image, the image area of the speaker can be roughly distinguished;

Step 2. Positioning process

(1) Calculate portrait-shadow edge distance

Since the image of the speaker and its shadow are likely to be superimposed on the image of the CCD video acquisition module, the deviation between the two can only be obtained by calculating the average distance between the corresponding edges;

(2) Calculate the overall width of the portrait-shadow

Detect the overall area of the segmented speaker image and its shadow, and obtain the overall width of the two;

(3) Calculating the distance between the portrait and the edge of the projected field

Calculate the distance between the speaker image and the adjacent edge of the projected content area;

(4) Calculation ratio

Assuming that the projector is parallel to the screen, even if this requirement is not met, the inclination of the projector is negligible under normal use. Two similar triangles are formed between the projector distance, the distance between the CCD module and the bulb, and according to the theorem of similar triangles, the projection content corresponding to the portrait-shadow edge distance, portrait-shadow overall width, and portrait-projection field edge distance can be calculated s position;

(5) Shadow image cutout

For the speaker image, the shadow is filled with black in the relative area of the projected content. Although the projector uses a bright bulb to illuminate the projected content, an image is formed on the screen, so even if the black content is projected, there is still projected light, but the projection The light intensity is greatly reduced, and the weakened projected light will not cause strong light damage to the eyes of the speaker.

The beneficial technical effects of the present invention are: adopting the method of installing a CCD video acquisition module in the projector machine makes image processing and data communication more convenient, the projector is easy to install, does not need large-scale wiring installation, and realizes the technical effect of instant use , In addition, by distinguishing the image of the speaker in front of the projector and the shadow position projected on the screen, and then estimating the relative position of the speaker, rapid positioning can be achieved, and when the position of the speaker changes, its shadow on the screen The position also changes, and the edge positions of the two will change immediately, thereby triggering a new positioning operation. There is no disadvantage of staying in multiple shadows, the algorithm is simple, and it has the advantages of simple operation, low hardware system requirements, low cost, and wide application. , with good economic benefits and market prospects.

Description of drawings

Fig. 1 is a kind of imaging structural diagram of the anti-glare method based on shadow image cutout algorithm of the present invention;

Fig. 2 is a kind of projector structural diagram of the anti-glare method based on shadow image cutout algorithm of the present invention;

Fig. 3 is a working principle diagram of an anti-glare method based on a shadow image cutout algorithm in the present invention.

Detailed ways

Referring to the accompanying drawings, the present invention will be described in detail below.

As shown in Figure 1, the common imaging structure of the present invention includes: projector (100), bulb (101), CCD video capture module (102), speaker (103), curtain (104), shadow (105). Wherein the CCD video acquisition module (102) is installed on the projector body.

As shown in Figure 2, the included parts of the projector (100): the light bulb (101) and the CCD video acquisition module (102) are installed on the same work surface, and the distance between the CCD module and the light bulb between the two centers is d( 200) known.

As shown in Figure 3, the lecturer (103) blocks in front of the curtain and enters the scope of the projected content area (301) of the audience, so the bulb (101) of the projector (100) leaves it on the curtain (104) shadow(105);

From the perspective of the CCD video acquisition module (102) of the projector (100), the speaker image (300) is also "projected" on the curtain (104), and forms a partial superposition with the shadow (105). In fact, The shadow (105) and the speaker image (300) in the figure should be on the same plane as the curtain (104), and for clarity of illustration, the front and back are staggered and marked in the figure;

Therefore, the detected portrait-shadow edge distance s (302), portrait-shadow overall width t (303), portrait-projected field edge distance p (304) has just determined its relative position on the screen;

Two similar triangles formed by portrait-shadow edge distance s(302), speaker-screen distance r1(305), CCD module-bulb distance d(200), speaker-projector distance r2(306), can be calculated The portrait-shadow edge spacing s (302), the portrait-shadow overall width t (303), and the portrait-projection field edge spacing p (304) respectively correspond to the positions of the projected content, so as to achieve the purpose of shadow image removal;

It can be seen that when the position of the speaker (103) moves, its shadow (105) on the curtain (104) changes immediately, thus triggering a new positioning. This method has good tracking performance, and there is no curtain (104) Disadvantages of stuck shadows on multiple blocks.

Claims (1)

1. An anti-glare method based on a shadow image cutout algorithm, adopting a CCD video acquisition module installed on the projector body to obtain the image of the lecturer in front of the projector and its shadow on the screen, its technical characteristics are as follows: Distinguish the image of the speaker in front of the projector and its shadow position projected on the screen, and then estimate the relative position of the speaker to realize the shadow image removal algorithm. The specific implementation steps include: Step 1. Image recognition process (1) Get the background image Since the projected content area of the whole field generally changes less, and the non-content area outside the curtain also changes less, it can be regarded as a background image, and a smooth and filtered background image can be obtained by repeatedly superimposing multiple frames of video images; (2) Divide to obtain the projected content area of the audience Due to the high brightness of the projected content area of the whole field, and the low brightness of the non-content area outside the screen, the rectangular area of the projected content area of the whole field can be segmented by calculating the brightness gradient map of the background image; (3) Identify shadows Under normal circumstances, the shadow on the screen is relatively easy to detect, and the shadow area can be directly distinguished by setting the brightness threshold; (4) Detect the speaker image area Through the difference operation with the background image, the image area of the speaker can be roughly distinguished; Step 2. Positioning process (1) Calculate portrait-shadow edge distance Since the image of the speaker and its shadow are likely to be superimposed on the image of the CCD video acquisition module, the deviation between the two can only be obtained by calculating the average distance between the corresponding edges; (2) Calculate the overall width of the portrait-shadow Detect the overall area of the segmented speaker image and its shadow, and obtain the overall width of the two; (3) Calculating the distance between the portrait and the edge of the projected field Calculate the distance between the speaker image and the adjacent edge of the projected content area; (4) Calculation ratio Assuming that the projector is parallel to the screen, even if this requirement is not met, the inclination of the projector is negligible under normal use. Two similar triangles are formed between the projector distance, the distance between the CCD module and the bulb, and according to the theorem of similar triangles, the projection content corresponding to the portrait-shadow edge distance, portrait-shadow overall width, and portrait-projection field edge distance can be calculated s position; (5) Shadow image cutout For the speaker image, the shadow is filled with black in the relative area of the projected content. Although the projector uses a bright bulb to illuminate the projected content, an image is formed on the screen, so even if the black content is projected, there is still projected light, but the projection The light intensity is greatly reduced, and the weakened projected light will not cause strong light damage to the eyes of the speaker.

Images