TWI657410B - Method and image processing system of image angle detection - Google Patents

Abstract

A method for image angle detection and an image processing system thereof, the method is applicable to an image processing system and includes obtaining a source image, and obtaining a source image corresponding to image compression or image decompression for a source image The first feature value is used to perform feature detection using the first feature value, and then the skew angle of the source image is calculated according to the result of the feature detection.

Description

Image angle detection method and image processing system thereof

The present invention relates to a technique for image processing, and more particularly to a method for image angle detection and an image processing system thereof.

Image scanners mainly analyze and convert photos, printed documents, handwritten documents and other objects into digital images by optical scanning. In the scanning process, the scanned images are often caused by human factors or the device itself. The image is skewed. Generally, the method of correcting image skew can only perform the operation of the entire image by using an algorithm after obtaining a complete scanned image. If the image is large, in addition to waiting for the image data to be completely obtained, it is necessary to start the calibration. It also requires more computing time and computing resources. Therefore, the image angle correction solution is not widely used in the low-end printing on the market. In the table or scanner.

In view of the above, the present invention provides a method for image angle detection and an image processing system thereof, which can complete image angle detection with a small amount of data, system computing resources, and time.

In an embodiment of the invention, the method is applicable to an image processing system and includes acquiring a source image, and obtaining a plurality of first images corresponding to the source image during image compression or image decompression for the source image. The feature value is subjected to feature detection using the first feature value, and the skew angle of the source image is calculated according to the result of the feature detection.

In an embodiment of the invention, the image processing system includes a memory and a processor, wherein the memory is coupled to the processor. Memory is used to store images and data. The processor is configured to obtain the source image, and obtain a plurality of first feature values corresponding to the source image in the process of image compression or image decompression for the source image, and perform feature detection using the first feature value, and according to the feature As a result of the detection, the skew angle of the source image is calculated.

The above described features and advantages of the invention will be apparent from the following description.

The components of the present invention will be described in detail in the following description in conjunction with the accompanying drawings. These examples are only a part of the invention and do not disclose all of the embodiments of the invention. Rather, these embodiments are merely examples of methods and systems within the scope of the patent application of the present invention.

1 is a block diagram of an image processing system according to an embodiment of the invention, but is for convenience of description and is not intended to limit the present invention. First, Figure 1 first introduces all the components and configuration relationships in the image processing system. The detailed functions will be disclosed in conjunction with Figure 2.

Referring to FIG. 1 , the image processing system 100 includes a memory 110 and a processor 120 , wherein the memory 110 is coupled to the processor 120 . In an embodiment, the image processing system 100 can be a computer system with image processing functions such as a personal computer, a notebook computer, a tablet computer, a workstation, a server computer, a large computer, and an industrial computer. In another embodiment, the image processing system 100 may include the above computer system and a scanner connected to it wired or wirelessly, or a scanning system, a printer, a multifunction machine, and the like that have both scanning and image processing functions. The device, the invention is not limited thereto.

The memory 110 is used for storing images, programs, and the like, and may be, for example, any type of fixed or removable random access memory (RAM), read-only memory (ROM). ), flash memory, hard disk or other similar device, integrated circuit, and combinations thereof.

The processor 120 is configured to control operations between components of the image processing system 100, which may be, for example, a central processing unit (CPU), or other programmable general purpose or special purpose microprocessors. ), digital signal processor (DSP), graphics processing unit (GPU), programmable controller, application specific integrated circuits (ASIC), programmable logic devices (programmable logic device, PLD), other similar devices, or a combination of the above.

FIG. 2 is a flowchart of a method for image angle detection according to an embodiment of the invention, and the method flow in this embodiment may be implemented by components of the image processing system 100.

Referring to FIG. 1 and FIG. 2 simultaneously, first, the processor 120 of the image processing system 100 receives the source image (step S202), and performs a process of image compression or image decompression for the source image (step S210). Here, the processor 120 determines whether to perform image compression or image decompression depending on the type of the source image. For example, if the source image is a raw image generated by the scanner, it is often required to be compressed into a compressed image such as a JPEG format based on its larger data amount to save storage space of the memory 110. Assuming that the source image is a compressed image received via the network, it needs to be decompressed for presentation on a screen (not shown).

When the processor 120 performs image compression or image decompression on the source image, a plurality of first feature values corresponding to the source image are obtained (step S212). The first characteristic value here is that the processor 120 has intermediate information generated by the conversion between the spatial domain and the frequency domain in the process of performing image compression or image decompression, so the processor 120 does not need to perform additional operations. This mediation data can represent most of the image features of the source image.

In the case of general image compression, a frequency separation process is performed to separate high frequency values from low frequency values in the image to achieve high compression ratio. The first characteristic value used in this embodiment is a low frequency value. . In this embodiment, when performing image compression on the source image, the processor 120 may first divide the image into a plurality of square blocks (for example, 8×8, 16×16) that are the same size and do not overlap, and then Each block performs a discrete cosine transform (DCT), and each block will generate a DCT coefficient matrix. In the case of an 8×8 block, its DCT coefficient matrix will have one DC coefficient (DC value) representing low frequency and 63 AC coefficients (AC value) representing high frequency. Here, all DC values corresponding to the source image will constitute a low frequency image, and the low frequency image is sufficient to describe most of the image features of the source image. Therefore, the processor 120 will retrieve the DC value of each block as the aforementioned "first feature value". It should be noted that, in other embodiments, the processor 120 may use a discrete wavelet transform, a Fourier transform, or the like to obtain the first feature value in a manner of band separation. This limit. It should be apparent to those skilled in the art that the processor 120 may perform color conversion and sampling before dividing the source image into a plurality of blocks. In addition, after the DCT conversion is performed for each block, the processor 120 will continue to perform processes such as quantization and entropy encoding, and step S212 does not affect the image compression process.

On the other hand, in the case of the general image decompression, in the embodiment, the processor 120 may first perform entropy decoding, inverse quantization, and the like on the source image to obtain the respective blocks. The DC value is used as the "first characteristic value" described above. Similarly, the processor 120 will continuously perform procedures such as inverse discrete cosine transform (IDCT), unsampling, and inverse color transformation, and step S212 will not affect the image solution. Compression process.

After obtaining the first feature value, the processor 120 performs feature detection using the first feature value (step S214), and calculates a skew angle of the source image according to the result of the feature detection (step S216). In other words, the processor 120 can perform image detection by using the source image for image compression or mediation data generated during image decompression to obtain the skew angle of the source image. Here, the processor 120 may detect the first feature by using a line detection algorithm such as a Hough transform, a Radon transform, or a regression analysis. The linear feature in the intermediate image formed by the value. In this embodiment, the processor 120 uses the low frequency image composed of the DC value to perform the line detection, and then the angle between the detected straight line segment and the horizontal direction or the vertical direction, that is, the source image can be obtained. Skew angle. Therefore, the processor 120 does not need to wait for image compression or image decompression, that is, the low-frequency image with a small amount of data can be used to calculate the skew angle of the source image. For the sake of convenience, the following embodiments will explain the application scenario of FIG. 2 in the future.

FIG. 3 is a flow chart of an application scenario of a method for image angle detection according to an embodiment of the invention. The method flow in this embodiment can be implemented by each component of the image processing system 100, and the scanned image generated by the scanner is used as a source image.

Referring to FIG. 1 and FIG. 3 simultaneously, first, the processor 120 of the image processing system 100 performs image compression on the scanned image Img30 (step S310) to be compressed into, for example, a progressive JPEG format or a baseline. Compressed image in JPEG format. During the image compression process, the processor 120 captures the DC value generated by the DCT conversion (step S312) to perform line detection (step S314), thereby obtaining the skew angle of the scanned image Img30 (step S316), wherein For details of the steps S310, S312, S314, and S316, refer to the foregoing embodiment, and details are not described herein again. Incidentally, when the scanned image is a file file, the majority of each line of text is arranged in a straight line, and the table and the image in the file have a straight line border, so the processor 120 can accurately calculate the skew of the scanned image. angle.

Then, after the processor 120 compresses the scanned image, the information about the skew angle is written into the compressed image to generate a compressed image Img31 having angle information. For EXIF information of the JPEG image, the processor 120 may, for example, write the skew angle to the UserComment tag (label number 0x9286) for storing the user's annotation. Next, the processor 120 performs image decompression on the compressed image Img31 (step S320) to generate a decompressed image Img32. At the same time, the processor 120 will read the angle information in the compressed image Img31 (step S322) to perform skew correction (deskew) on the decompressed image Img32 according to the angle information, and generate the corrected decompressed image Img33. In other words, step S320 and step S322 can be understood to be performed simultaneously, and the read angle information of step S322 is not continued until image decompression is completed in step S320.

The processor 120 of the embodiment estimates the offset angle of the scanned image Img30 by using the low frequency image corresponding to the scanned image Img30 in the process of image compression for the scanned image Img30, so that the data amount can be reduced by a small amount of data. The accuracy of the offset angle analysis with the complete image can reduce the consumption and time of the system computing resources. In addition, since the skew angle has been written in the compressed image Img31, the processor 120 does not need to recalculate the offset angle each time the compressed image Img31 is turned on.

It is worth mentioning that, assuming that the scanned image Img30 is a raw RGB image of a plain text file, the resolution has 300 dpi (2480×3508 pixels), and the low frequency image composed of the DC value is (310×439 pixels). Taking the 64-bit (x64) version of the Windows 7 operating system as an example, if the Intel Core i7-3520M 2.9 GHz is used as the processor 120, the conventional process requires 672 ms to obtain a complete decompressed image. After that, the line detection is performed for 33 ms to obtain the offset angle, and the flow of FIG. 3 only takes 156 ms to obtain the offset angle. Therefore, the image processing system 100 can be implemented as a medium and low-priced photocopier, printer, scanner, or multifunction printer to provide an option to correct existing documents, thereby reducing the cost of copying the document paper and ink.

In addition to correcting the aforementioned scanned images, the image processing system 100 can be used to correct general images to facilitate subsequent applications. Specifically, FIG. 4 is a flow chart of an application scenario of a method for image angle detection according to an embodiment of the invention. The method flow in this embodiment can be implemented by various components of the image processing system 100, and the compressed image received from the network is used as a source image, which is reasonably assumed here. Compressed images do not have angular information.

Referring to FIG. 1 and FIG. 4 simultaneously, first, the processor 120 of the image processing system 100 performs image decompression on the compressed image Img40 having no angle information (step S410). In the process of performing image compression, the processor 120 will take a DC value after performing the inverse quantization process (ie, before the IDCT conversion) (step S412) to perform line detection (step S414), thereby obtaining the skew of the compressed image Img40. Angle (step S416). For details of the steps S410, S412, S414, and S416, refer to the foregoing embodiment, and details are not described herein again. After the processor 120 decompresses the compressed image Img40, the decompressed image Img41 is generated, and the decompressed image Img41 is further corrected according to the skew angle, and the corrected decompressed image Img42 is generated to increase the use. The more you read the comfort.

In summary, the image angle detection method and the image processing system thereof according to the present invention use image representing most of the source image during image compression or image decompression for the source image. The feature is also related to the mediation data generated by the conversion between the spatial domain and the frequency domain for image angle detection. Therefore, the image processing system of the present invention does not need to perform additional operations, except that the accuracy of the offset angle analysis can be achieved with a small amount of data, and the consumption and time of the system computing resources can be reduced, so that the image can be implemented in the market. High- and low-level related products to enhance the consumer's better user experience.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention, and any one of ordinary skill in the art can make some changes and refinements without departing from the spirit and scope of the present invention. The scope of the invention is defined by the scope of the appended claims.

100‧‧‧Image Processing System

110‧‧‧ memory

120‧‧‧ processor

S202, S212, S214, S216, S310, S312, S314, S316, S320, S322, S410, S412, S414, S416‧‧ steps

Img30, Img31, Img32, Img33, Img40, Img41, Img42‧‧ images

FIG. 1 is a block diagram of an image processing system according to an embodiment of the invention. 2 is a flow chart of a method for image angle detection according to an embodiment of the invention. FIG. 3 is a flow chart of an application scenario of a method for image angle detection according to an embodiment of the invention. FIG. 4 is a flow chart of an application scenario of a method for image angle detection according to another embodiment of the invention.

Claims (14)

An image angle detection method is applicable to an image processing system, and the method includes: obtaining a source image; and performing image compression or image decompression on the source image: obtaining a plurality of the corresponding images of the source image a feature value; performing feature detection by using the first feature values; and calculating a skew angle of the source image according to the result of the feature detection, wherein the source image is obtained during the image compression process for the source image The step of the corresponding first feature values includes: obtaining a plurality of low frequency values generated by performing spatial domain to frequency domain transform coding on the source image as the first feature values. The method of claim 1, wherein the first feature values are associated with a low frequency signal of the source image. The method of claim 1, wherein the source image is a scanned image. The method of claim 1, wherein the conversion code is a discrete cosine transform, and each of the low frequency values is a DC coefficient (DC value). The method of claim 1, wherein the step of performing feature detection using the first feature values in the process of performing the image compression on the source image comprises: Line detection is performed using the first feature values to detect line segments in the source image. The method of claim 1, wherein after performing the image compression process on the source image, the method further comprises: generating a compressed image of the source image; and writing related information of the skew angle to The compressed image. The method of claim 6, further comprising: when the image is decompressed for the compressed image, reading the related information of the skew angle; and generating image decompression for the compressed image When the image is decompressed, the decompressed image is corrected based on the correlation information of the skew angle to generate a corrected decompressed image. The method of claim 1, wherein the source image is a compressed image. The method of claim 8, wherein the step of obtaining the first feature values of the source image during the image decompression of the source image comprises: obtaining a plurality of low frequencies of the compressed image Values are used as the first feature values. The method of claim 9, wherein each of the low frequency values is a direct current coefficient (DC value). The method of claim 9, wherein the step of performing feature detection according to the first feature values during the image decompression of the source image to generate the detection result comprises: utilizing the The first feature value is linearly detected to detect a line segment in the source image. The method of claim 8, wherein after performing the image decompression process on the source image, the method further comprises: generating a decompressed image of the source image; and correcting the angle according to the skew angle The image is decompressed to produce a corrected decompressed image. The method of claim 1, wherein the image compression process is a JPEG compression process, and the image decompression process is a JPEG decompression process. An image processing system includes: a memory for storing images and data; a processor coupled to the memory for acquiring a source image, and performing image compression or image decompression for the source image, And acquiring a plurality of first feature values corresponding to the source image, performing feature detection by using the first feature values, and calculating a skew angle of the source image according to the result of the feature detection, where the source image is The step of obtaining the first feature values corresponding to the source image during the image compression process comprises: the processor is configured to obtain a spatial domain to a frequency domain for the source image A plurality of low frequency values generated by the encoding are changed as the first characteristic values.