Dynamic image projection system and projection method thereof

Abstract

The invention discloses a dynamic image projection system, and the system comprises a video pre-reading module which is in communication connection with a video input port, and is used for pre-reading a to-be-projected video, and carrying out the temporary storage and preprocessing of the pre-read video; the video correction module is in communication connection with the video pre-reading module and is used for correcting the video before projection; the projection module is in communication connection with the video correction module and is used for projecting a video; and the projection image acquisition module is in communication connection with the video correction module and is used for acquiring projected image information and feeding back the acquired image information to the video correction module. According to the method, the defects in the prior art can be overcome, and the accuracy of image analysis is ensured while the feedback speed is increased.

Description

Technical Field

[0001] This invention relates to the field of multimedia technology, and in particular to a dynamic image projection system and its projection method. Background Technology

[0002] Dynamic image projection refers to displaying dynamic video images through projection. Unlike static projection, dynamic projection requires projecting multiple video frames sequentially within a very short time. Current technologies for controlling dynamic projection primarily employ open-loop control, meaning the projection system is adjusted and calibrated based on preset values ​​before projection. Some research has attempted to implement feedback adjustment during projection; however, due to the rapid frame changes in dynamic projection, ordinary feedback adjustment mechanisms cannot adapt to such high-speed input changes, resulting in significant feedback lag. Furthermore, simplifying the feedback adjustment mechanism to improve feedback speed directly impacts the accuracy of image analysis results. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a dynamic image projection system and projection method that can overcome the shortcomings of the prior art and improve the feedback speed while ensuring the accuracy of image analysis.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows.

[0005] A dynamic image projection system, comprising, The video pre-reading module communicates with the video input port and is used to pre-read the video to be projected, and to temporarily store and preprocess the pre-read video; The video correction module communicates with the video pre-reading module and is used to correct the video before projection. The projection module, which communicates with the video correction module, is used to project video. The projection image acquisition module is connected to the video correction module and is used to acquire the projected image information and feed the acquired image information back to the video correction module.

[0006] Preferably, the video pre-reading module has a video temporary storage space of more than 200M.

[0007] A projection method for the above-mentioned dynamic image system includes the following steps: A. The video pre-reading module pre-reads the video to be projected through the video input port and temporarily stores and preprocesses the pre-read video; B. The video correction module corrects the video before projection based on the image information fed back from the projection image acquisition module. C. The projection module projects the corrected video image. D. The projection image acquisition module acquires the projected image information and feeds the acquired image information back to the video correction module.

[0008] Preferably, step A involves preprocessing the pre-read video, including the following steps. The video frames are traversed in time sequence, and the feature matrix of each video frame is extracted. The video frames are then grouped in time sequence according to the linear correlation of the feature matrices. Video frames in the same group are sequential in time, and the tag matrix is ​​extracted using the feature matrix set of the video frames in the same group.

[0009] Preferably, the steps for extracting the video frame feature matrix include: Traverse the pixels of the video frame, calculate the gradient direction of pixel grayscale change, mark the closed path with the minimum average rate of change of pixel grayscale along the normal direction of the pixel grayscale change gradient, traverse the closed path, and collect the pixel grayscale values ​​of a set length along the pixel grayscale change gradient direction as the feature matrix of the video frame. The set length is related to the total number of pixels in the video frame and the variance of pixel grayscale. The larger the total number of pixels, the longer the set length. The larger the variance of pixel grayscale, the longer the set length.

[0010] Preferably, the step of extracting the marker matrix includes, The feature matrix is ​​divided into several sub-matrices, and the position coordinates of each sub-matrix in the feature matrix are assigned. The sub-matrices with the same position coordinates in different feature matrices in the feature matrix set are compared. If the similarity of the sub-matrices at the same position coordinates exceeds a set threshold, the sub-matrix at the above position coordinates is set as the label matrix. The rank of the sub-matrix is ​​greater than 1.

[0011] Preferably, step B, which involves correcting the video before projection, includes the following steps: Extract the label matrix of the image information fed back by the projection image acquisition module, compare the label matrix with the label matrix of the corresponding video frame, fit the tangential distortion of the current projection based on the coordinate deviation of the label matrix, fit the brightness drift of the current projection based on the standard deviation of the gray values ​​of the label matrix, and correct the current projection video frame based on the fitted tangential distortion and brightness drift.

[0012] As a preferred method, the fitting method for the tangential distortion variable is as follows: Record the coordinate deviations of all the marked matrices. Each coordinate deviation generates a deviation vector. Map the deviation vectors to the coordinate system. Assign weight values ​​to the deviation vectors according to their clustering degree. The farther the deviation vector is from other deviation vectors, the higher its weight value. Then, the deviation vectors are weighted and merged to obtain the total deviation vector. Determine the distortion angle and distortion magnitude of the tangential distortion variable based on the angle and magnitude of the total deviation vector.

[0013] As a preferred method, the fitting method for brightness drift is as follows: The weight values ​​of the deviation vector are assigned to the corresponding label matrix. Then, the average gray values ​​of the label matrix are weighted, and the standard deviation of the gray values ​​after weighting all label matrices is calculated. The obtained standard deviation of gray values ​​is proportional to the brightness offset.

[0014] The beneficial effects of adopting the above technical solution are as follows: By optimizing the video frame processing method, the present invention extracts the key marker matrix for comparison feedback, which greatly reduces the comparison calculation while ensuring the accuracy of the comparison, thereby effectively improving the feedback speed. Attached Figure Description

[0015] Figure 1 This is a structural diagram of a specific embodiment of the present invention. Detailed Implementation

[0016] Currently, the technical bottleneck of dynamic projection feedback mechanisms lies in how to quickly and accurately process large amounts of video frames. While increasing processing speed through hardware upgrades is theoretically feasible, it would significantly increase hardware costs, making it impractical for the market. Therefore, we approached the issue from the perspective of video frames themselves, finding a completely new feedback control strategy, which includes the following steps: A. The video pre-reading module 1 pre-reads the video to be projected through the video input port, and temporarily stores and pre-processes the pre-read video; B. The video correction module 2 corrects the video before projection based on the image information fed back from the projection image acquisition module 4. C. Projection module 3 projects the corrected video image; D. The projection image acquisition module 4 acquires the projected image information and feeds the acquired image information back to the video correction module 2.

[0017] The key lies in the corrections we make to the video before projection. Specifically, this includes: The video frames are traversed sequentially, and the feature matrix of each video frame is extracted. The video frames are then grouped sequentially according to the linear correlation of the feature matrices, with video frames within the same group being sequentially continuous. A label matrix is ​​extracted using the feature matrix set of the same group. The label matrix of the image information fed back from the projection image acquisition module 4 is extracted. This label matrix is ​​compared with the label matrix of the corresponding video frame. The tangential distortion of the current projection is fitted based on the coordinate deviation of the label matrix, and the brightness drift of the current projection is fitted based on the standard deviation of the grayscale values ​​of the label matrix. The current projection video frame is then corrected based on the fitted tangential distortion and brightness drift. Through this process, we extract the label matrix as the object for subsequent comparison feedback, which reduces the amount of comparison data while ensuring accuracy. To ensure that the label matrix reflects the state of the video frame to the greatest extent, we specifically designed a label matrix extraction method, as follows: The process iterates through the pixels of the video frame, calculates the gradient direction of pixel grayscale changes, and marks the closed path with the minimum average rate of change of pixel grayscale along the normal direction of the pixel grayscale gradient. It then iterates through this closed path, collecting pixel grayscale values ​​of a predetermined length along the gradient direction. This predetermined length is used as the feature matrix of the video frame. The predetermined length is related to the total number of pixels in the video frame and the variance of pixel grayscale values; a larger total number of pixels and a larger variance in pixel grayscale values ​​both require a longer predetermined length. The feature matrix is ​​divided into several sub-matrices, and each sub-matrix is ​​assigned its position coordinates within the feature matrix. Sub-matrices with the same position coordinates in different feature matrices within the feature matrix set are compared. If the similarity of sub-matrices at the same position coordinates exceeds a predetermined threshold, then the sub-matrix at those position coordinates is designated as a marker matrix. The rank of each sub-matrix is ​​greater than 1.

[0018] The design requirement for the marker matrix is ​​to encompass most of the key features of the video frame while ensuring its size is not too large. To extract the key features of the video frame, we first constructed a feature matrix. The closed path with the minimum average rate of change of pixel grayscale along the normal direction of the pixel grayscale gradient contains the main trends of pixel grayscale changes in the video frame, while reducing the rate of repeated sampling. The marker matrix is ​​a submatrix of the feature matrix. Using the marker matrix for alignment not only reduces the dimensionality of the feature matrix and improves alignment efficiency, but more importantly, it introduces the concept of "position coordinates" of the marker matrix, thus creating conditions for subsequent alignment.

[0019] The fitting method for the tangential distortion variable is as follows: Record the coordinate deviations of all the marked matrices. Each coordinate deviation generates a deviation vector. Map the deviation vectors to the coordinate system. Assign weight values ​​to the deviation vectors according to their clustering degree. The farther the deviation vector is from other deviation vectors, the higher its weight value. Then, the deviation vectors are weighted and merged to obtain the total deviation vector. Determine the distortion angle and distortion magnitude of the tangential distortion variable based on the angle and magnitude of the total deviation vector.

[0020] The fitting method for brightness drift is as follows: The weight values ​​of the deviation vector are assigned to the corresponding label matrix. Then, the average gray values ​​of the label matrix are weighted, and the standard deviation of the gray values ​​after weighting all label matrices is calculated. The obtained standard deviation of gray values ​​is proportional to the brightness offset.

[0021] Tangential distortion and brightness drift are the two most common defects in projection systems. Our designed marker matrix can be accurately weighted and fitted using its position coordinates, and the defect comparison results of the entire video frame can be obtained by using the local features of the video frame it carries.

[0022] This invention achieves fast and accurate feedback control by creating a marker matrix and fitting the two main defect characteristics of tangent distortion and brightness offset in the projection process.

[0023] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this invention, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0024] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A dynamic image projection system, characterized in that: include, The video pre-reading module (1) is connected to the video input port for pre-reading the video to be projected and temporarily storing and pre-processing the pre-read video; The video correction module (2) is connected to the video pre-reading module (1) and is used to correct the video before projection. The projection module (3) is connected in communication with the video correction module (2) and is used to project video; The projection image acquisition module (4) is connected to the video correction module (2) for acquiring the projected image information and feeding back the acquired image information to the video correction module (2).

2. The dynamic image projection system according to claim 1, characterized in that: The video pre-reading module (1) has a video temporary storage space of more than 200M.

3. A projection method for a dynamic image system as described in claim 1 or 2, characterized in that... Includes the following steps: A. Video pre-reading module (1) pre-reads the video to be projected through the video input port, and temporarily stores and pre-processes the pre-read video; B. The video correction module (2) corrects the video before projection based on the image information fed back by the projection image acquisition module (4); C. The projection module (3) projects the corrected video image; D. Projection image acquisition module (4) acquires the projected image information and feeds the acquired image information back to the video correction module (2).

4. The projection method of the dynamic imaging system according to claim 3, characterized in that: Step A involves preprocessing the pre-read video, including the following steps: The video frames are traversed in time sequence, and the feature matrix of each video frame is extracted. The video frames are then grouped in time sequence according to the linear correlation of the feature matrices. Video frames in the same group are sequential in time, and the tag matrix is ​​extracted using the feature matrix set of the video frames in the same group.

5. The projection method of the dynamic imaging system according to claim 4, characterized in that: The steps for extracting the feature matrix of video frames include: Traverse the pixels of the video frame, calculate the gradient direction of pixel grayscale change, mark the closed path with the minimum average rate of change of pixel grayscale along the normal direction of the pixel grayscale change gradient, traverse the closed path, and collect the pixel grayscale values ​​of a set length along the pixel grayscale change gradient direction as the feature matrix of the video frame. The set length is related to the total number of pixels in the video frame and the variance of pixel grayscale. The larger the total number of pixels, the longer the set length. The larger the variance of pixel grayscale, the longer the set length.

6. The projection method of the dynamic imaging system according to claim 5, characterized in that: The steps for extracting the tag matrix include: The feature matrix is ​​divided into several sub-matrices, and the position coordinates of each sub-matrix in the feature matrix are assigned. The sub-matrices with the same position coordinates in different feature matrices in the feature matrix set are compared. If the similarity of the sub-matrices at the same position coordinates exceeds a set threshold, the sub-matrix at the above position coordinates is set as the label matrix. The rank of the sub-matrix is ​​greater than 1.

7. The projection method of the dynamic imaging system according to claim 6, characterized in that: Step B, the pre-projection correction of the video, includes the following steps: Extract the label matrix of the image information fed back by the projection image acquisition module (4), compare the above label matrix with the label matrix of the corresponding video frame, fit the tangential distortion of the current projection according to the coordinate deviation of the label matrix, fit the brightness drift of the current projection according to the standard deviation of the gray value of the label matrix, and correct the current projection video frame according to the fitted tangential distortion and brightness drift.

8. The projection method of the dynamic image system according to claim 7, characterized in that: The fitting method for the tangential distortion variable is as follows: Record the coordinate deviations of all the marked matrices. Each coordinate deviation generates a deviation vector. Map the deviation vectors to the coordinate system. Assign weight values ​​to the deviation vectors according to their clustering degree. The farther the deviation vector is from other deviation vectors, the higher its weight value. Then, the deviation vectors are weighted and merged to obtain the total deviation vector. Determine the distortion angle and distortion magnitude of the tangential distortion variable based on the angle and magnitude of the total deviation vector.

9. The projection method of the dynamic image system according to claim 8, characterized in that: The fitting method for brightness drift is as follows: The weight values ​​of the deviation vector are assigned to the corresponding label matrix. Then, the average gray values ​​of the label matrix are weighted, and the standard deviation of the gray values ​​after weighting all label matrices is calculated. The obtained standard deviation of gray values ​​is proportional to the brightness offset.

Images