TWI520603B - Method for processing bad pixel and image processing apparatus thereof - Google Patents

Description

Bad pixel processing method and image processing device thereof

The present invention relates to an image processing technique, and more particularly to a bad pixel processing method and an image processing apparatus thereof for images having different exposure times.

The "dynamic range" refers to the range or ratio of the maximum brightness value and the minimum brightness value in the picture. For photography, the dynamic range can be further divided into "dynamic range of image sensor" and "dynamic range of scene". Among them, the dynamic range of the image sensor refers to the range in which the photosensitive element can accept the change in brightness. The dynamic range of the scene refers to the range of brightness differences in the shooting scene, that is, the difference between the brightest and darkest areas in the scene.

When the dynamic range of the scene is larger than the dynamic range of the image sensor, it means that there are extreme bright and dark parts in the shooting scene, which exceeds the color gradation that the photosensitive element can record, so the black or white area will appear in the photo. Piece. In order to overcome this drawback, the High Dynamic Range (HDR) image sensor transmits image processing technology, so that the dynamic range of the processed image is larger than the dynamic range provided by a single image captured by a general camera.

One of the operating modes of the high dynamic range image sensor is to generate images of two long exposures and two short exposures in the same picture. A single image obtained by continuous alternate shooting with long exposure and short exposure will produce a partial exposure time different from that of other pixels, but the existing dead pixel correction technique is based on the assumption that a single image has only a single exposure time. That is to say, the use of long exposure and short exposure continuous alternate shooting of a single image does not directly apply the existing dead pixel correction technique. Accordingly, the above-mentioned bad point correction problem will affect the quality of high dynamic range images.

The present invention provides a bad pixel processing method and an image processing apparatus thereof, which can detect bad pixels and correct bad pixels when one image has different exposure times.

An embodiment of the present invention provides a bad pixel processing method for an image processing apparatus. The method includes: obtaining a plurality of pixels, wherein a plurality of first pixels of the pixels have a first exposure time, and a plurality of second pixels of the pixels have a second exposure time, the first exposure time being different from the second Exposure time, and the current pixel in the pixel belongs to a first exposure time; adjusting a color value of each second pixel to correspond to a first exposure time; determining whether the pixel is overexposed to generate a determination result; according to the first The pixel, the adjusted second pixel and the determination result determine whether the current pixel is a bad pixel; and if the current pixel is a bad pixel, correct the current pixel.

In one embodiment, the first exposure time is one of a short exposure and a long exposure, and the second exposure time is one of a short exposure and a long exposure, and the color value of each second pixel is adjusted as described above. Steps corresponding to the first exposure time The method includes: if the second exposure time is a short exposure, multiplying a color value of each second pixel by a gain, wherein the gain is calculated according to the first exposure time and the second exposure time; and if the second exposure time is a long exposure The color value of each second pixel is divided by the gain.

In an embodiment, the step of determining whether the pixel is overexposed to generate a determination result comprises: obtaining a test pixel in the pixel; and if the test pixel has a long exposure, determining whether a color value of the test pixel is greater than a first threshold And if the color value of the test pixel is greater than the first threshold, determining that the test pixel is overexposed; if the test pixel has a short exposure, determining whether the product of the color value of the test pixel multiplied by the gain is greater than a second threshold, and if the product Greater than the second threshold, determining that the test pixel is overexposed; and marking whether the test pixel is overexposed in the determination result.

In an embodiment, the plurality of third pixels in the pixel have the same channel as the current pixel, and the step of determining whether the current pixel is a bad pixel includes: determining, according to the determination result, whether each third pixel is overexposed; For each third pixel, if the corresponding third pixel is overexposed, a count value is updated; for each third pixel, if the corresponding third pixel is not overexposed, according to the color of the corresponding third pixel The difference between the value and the color value of the current pixel is used to update the count value; whether the count value is greater than the third critical value; and if the count value is not greater than the third critical value, it is determined that the current pixel is not a bad pixel.

In an embodiment, the step of updating the count value according to the difference between the color value of the corresponding third pixel and the color value of the current pixel includes: determining the fourth critical value and the fifth critical value according to the color value of the current pixel. Value; if the gap is greater than the fourth The threshold value is updated, and if the difference is less than the fifth threshold value, the count value is updated; and if the difference is between the fourth threshold value and the fifth threshold value, the count value is maintained unchanged.

In an embodiment, the step of determining whether the current pixel is a bad pixel further comprises: if the count value is greater than the third threshold, determining whether one of the third pixels is overexposed; and if the third pixel Any one of them is overexposed, and it is judged that the current pixel is a bad pixel.

In an embodiment, the step of modifying the current pixel comprises: correcting the current pixel according to the third pixel having the first exposure time.

In an embodiment, the step of modifying the current pixel based on the third pixel having the first exposure time is performed according to a weighted average algorithm.

Another embodiment of the present invention provides an image processing apparatus including an upsampling circuit, a detecting circuit, and a repairing circuit. The up sampling circuit is configured to obtain a plurality of pixels, wherein the plurality of first pixels of the pixels have a first exposure time, and the plurality of second pixels of the pixels have a second exposure time, the first exposure time being different from the first Two exposure times, and the current pixel in the pixel belongs to a first exposure time, wherein the up sampling circuit is configured to adjust a color value of each second pixel to correspond to the first exposure time, and determine whether the pixel is overexposed to generate a determination result . The detecting circuit is coupled to the up sampling circuit for determining whether the current pixel is a bad pixel according to the first pixel, the adjusted second pixel, and the determination result. The repair circuit is coupled to the detection circuit. If the current pixel is a bad pixel, the detection circuit is used to correct the current pixel.

In one embodiment, the first exposure time is one of a short exposure and a long exposure, and the second exposure time is one of a short exposure and a long exposure. The step of adjusting the color value of each second pixel to correspond to the first exposure time includes: if the second exposure time is a short exposure, the up sampling circuit multiplies the color value of each second pixel by a gain, wherein the gain It is calculated according to the first exposure time and the second exposure time; and if the second exposure time is a long exposure, the up sampling circuit divides the color value of each second pixel by the gain.

In an embodiment, the above-mentioned up sampling circuit determines whether the pixel is overexposed to generate a determination result, including: the up sampling circuit obtains the test pixel in the pixel; if the test pixel has a long exposure, the up sampling circuit determines the test Whether the color value of the pixel is greater than the first critical value, and if the color value of the test pixel is greater than the first critical value, the up sampling circuit determines that the test pixel is overexposed; if the test pixel has a short exposure, the up sampling circuit determines the color value of the test pixel Whether the product after multiplying the gain is greater than the second threshold, and if the product is greater than the second threshold, the up sampling circuit determines that the test pixel is overexposed; and the up sampling circuit marks whether the test pixel is overexposed in the determination result.

In an embodiment, the plurality of third pixels in the pixel have the same channel as the current pixel, and the detecting circuit determines whether the current pixel is a bad pixel, and the detecting circuit determines whether each third pixel is determined according to the determination result. For each third pixel, if the corresponding third pixel is overexposed, the detection circuit updates a count value; for each third pixel, if the corresponding third pixel is not overexposed, then the detector The measuring circuit updates the counting value according to the difference between the color value of the corresponding third pixel and the color value of the current pixel; the detecting circuit determines whether the counting value is greater than the third critical value; and if the counting value is not greater than the third critical value, The detection circuit determines that the current pixel is not For bad pixels.

In an embodiment, the detecting circuit updates the count value according to the difference between the color value of the corresponding third pixel and the color value of the current pixel, and the detecting circuit determines the first color according to the color value of the current pixel. The fourth critical value and the fifth critical value; if the difference is greater than the fourth critical value, the detecting circuit updates the counting value; if the difference is less than the fifth critical value, the detecting circuit updates the counting value; and if the difference is between the fourth critical value and the Between the five critical values, the detection circuit maintains the count value unchanged.

In one embodiment, if the count value is greater than the third threshold, the detecting circuit further determines whether one of the third pixels is overexposed, and if any one of the third pixels is overexposed, The detection circuit determines that the current pixel is a bad pixel.

In an embodiment, the repair circuit described above corrects the current pixel based on the third pixel having the first exposure time.

In an embodiment, the repair circuit described above performs the above-described operation of correcting the current pixel according to the third pixel having the first exposure time according to a weighted average algorithm.

Based on the above, the bad pixel processing method and the image processing apparatus proposed by the embodiments of the present invention adjust the color values of the first pixel/second pixel different from the exposure time of the current pixel to correspond to the exposure time of the current pixel. . Moreover, the determination result of whether the pixel is overexposed is used to determine whether the current pixel is a bad pixel, and further corrects the bad pixel. Thereby, it is possible to detect bad pixels when the images have different exposure times.

In order to make the above features and advantages of the present invention more apparent, the following is a special The embodiments are described in detail below in conjunction with the drawings.

100‧‧‧Image processing device

102‧‧‧ liter sampling circuit

104‧‧‧Detection circuit

106‧‧‧Repair circuit

200‧‧‧ original image

R1~R8‧‧‧ horizontal pixel column

W ₁ ~W ₃ ‧‧‧Operation block

LE‧‧‧Long exposure time

SE‧‧‧Short exposure time

G ₁ , B ₂ , G ₃ , B ₄ , G ₅ , R ₆ , G ₇ , R ₈ , G ₉ , R ₁₀ , G ₁₁ , B ₁₂ , G ₁₃ , B ₁₄ , B ₁₅ , R ₁₆ , G ₁₇ , R ₁₈ , G ₁₉ , R ₂₀ , G ₂₁ , B ₂₂ , G ₂₃ , B ₂₄ , G _25, G' ₁ , B' ₂ , G' ₃ , B' ₄ , G' ₅ , R' ₆ , G ' ₇ , R' ₈ , G' ₉ , R' ₁₀ , G' ₂₁ , B' ₂₂ , G' ₂₃ , B' ₂₄ , G' ₂₅ ‧ ‧ pixels

Ws‧‧‧Short exposure block

310‧‧‧Results

S401~S409, S501~S513‧‧‧Steps of bad pixel processing method

FIG. 1 is a block diagram of an image processing apparatus according to an embodiment of the invention.

FIG. 2 is a schematic diagram of an original image obtained by an image sensor using different exposure times according to an embodiment of the invention.

FIG. 3 is a schematic diagram of a process of performing block processing by an operation block according to an embodiment of the invention.

FIG. 4 is a flowchart of a bad pixel processing method according to an embodiment of the invention.

FIG. 5 is a flowchart of a bad pixel processing method according to an embodiment of the invention.

The existing bad pixel detection method is based on a process in which a full picture is a single exposure time image, and when the image to be processed becomes an image in which two long exposure lines and two short exposure lines alternate continuously, for example, For images produced by high dynamic range HDR sensors, this method is not directly applicable. Even if the defect detection is repaired for an area with a single exposure time, the nearest neighbor, the same exposure time, and the pixels of the same color are not close enough to the pixel to be detected, so the false positive or the undetectable will be increased. The probability of affecting the quality of the repair. In order to improve the probability of detecting bad pixels and repair quality, In terms of pixels, if the surrounding pixels of the pixel to be processed can be upsampling, and at the same time, which surrounding pixels are marked with overexposure, the pixel after performing the upsampling and the above-mentioned mark are used for bad pixel detection. Test and repair will accurately detect and repair bad pixels. The present invention is a bad pixel processing method and an image processing apparatus proposed based on the above viewpoint. In order to make the content of the present invention clearer, the following specific embodiments are examples of the invention that can be implemented.

FIG. 1 is a block diagram of an image processing apparatus according to an embodiment of the invention. Referring to FIG. 1 , the image processing apparatus 100 includes an up sampling circuit 102 , a detection circuit 104 , and a repair circuit 106 . The image processing device 100 is, for example, a digital camera, a Digital Single Lens Reflex (DSLR) camera, a Digital Video Camcorder (DVC), or the like. In another embodiment, the image processing apparatus 100 can be built in an electronic device such as a smart phone, a tablet, or a notebook computer, and is not limited to the above.

In this embodiment, the image processing device 100 can have an image sensor (not shown) for capturing and capturing images. The image sensor may include a lens, a photosensitive element, a shutter module, and the like. The photosensitive element is, for example, a Complementary Metal-Oxide Semiconductor (CMOS) element or other element. The shutter module is composed of, for example, a plurality of blades and a driving module, and can be used to control the exposure time of each of the photosensitive elements in line-by-line. It should be noted that the image sensor described above is based on the first exposure time (this embodiment is long as an example) The exposure time, referred to as long exposure) and the second exposure time (exemplified by short exposure time in this embodiment, short exposure) are successively interlaced to obtain an original image, wherein the long exposure time is different from the short exposure time. In one embodiment, the image sensor can control, for example, two horizontal pixel columns to take a long exposure for shooting, and two horizontal pixel columns to take a short exposure for shooting, which are sequentially performed alternately.

For example, FIG. 2 is a schematic diagram of an original image obtained by using an image sensor with different exposure times according to an embodiment of the invention. Referring to FIG. 2, the first horizontal pixel column R1 and the second horizontal pixel column R2 in the original image 200 adopt a long exposure LE; the third horizontal pixel column R3 and the fourth horizontal pixel column R4 adopt a short exposure SE; The horizontal pixel column R5 and the sixth horizontal pixel column R6 adopt a long exposure LE; the seventh horizontal pixel column R7 and the eighth horizontal pixel column R8 adopt a short exposure SE. The image sensor is not limited to the above, and in another embodiment, the image sensor can also perform long exposure shooting for a single horizontal pixel column and short exposure shooting for a double horizontal pixel column. Alternatively, the image sensor can perform short exposure on the two horizontal pixel columns after long exposure of the three horizontal pixel columns.

It is to be noted that the image processing apparatus 100 of the present embodiment performs bad pixel detection and repair, for example, in units of operating blocks. In detail, the operation block may be referred to as a window, and the size of the operation block is, for example, N ^* N pixels, where N is a positive integer, but the present invention does not limit the value of N. The size of the operation blocks W ₁ , W ₂ , and W ₃ in the original image 200 is set to, for example, 5 ^* 5 pixels. It should be noted that the operation block of the original image 200 is not limited to three, and is only used as an example for illustration. In practice, the operation block to which it belongs may be set for each output pixel. For each of the operational blocks, one pixel is produced by the image processing apparatus 100.

FIG. 3 is a schematic diagram of a process of performing block processing by the operation block W ₁ according to an embodiment of the invention. Referring to FIG 3, in the present embodiment, operations block W ₁ comprises 25 pixels, and wish to process pixel as a pixel G _13. The 25 pixels in the operation block W ₁ are bayer arrangements, and each pixel has a color value (or channel). For example, the pixels G ₁ , G ₃ , G _{5 ,} etc. have green color values; the pixels B ₂ , B _{4 ,} etc. have blue color values; and the pixels R ₆ , R ₈ , R _{10 ,} etc. have red color values. It should be noted that the Bell arrangement in FIG. 3 is only an example. In other embodiments, the operation block W ₁ may have other arrangements. For example, a first pixel operations block pixel array R1 W ₁ may be arranged starting from the red or blue pixels, the present invention is permitted. The pixels G ₁ , B ₂ , G ₃ , B ₄ , G ₅ , R ₆ , G ₇ , R ₈ , G ₉ , R ₁₀ , G ₂₁ , B ₂₂ , G ₂₃ , B ₂₄ in the operation block W ₁ and The pixel G ₂₅ has a long exposure time; and the pixels G ₁₁ , B ₁₂ , G ₁₃ , B ₁₄ , G ₁₅ , R ₁₆ , G ₁₇ , R ₁₈ , G ₁₉ and the pixel R ₂₀ have a short exposure time. Hereinafter, a pixel having a long exposure time is referred to as a long exposure pixel, and a pixel having a short exposure time is referred to as a short exposure pixel.

The functions of the upsampling circuit 102, the detecting circuit 104, and the repairing circuit 106 in the image processing apparatus 100 will be described below with reference to FIGS. 1 and 3. Referring to FIG. 1 and FIG. 3, the pixel G _{13 that} is to be processed in this embodiment is the current pixel G ₁₃ , and the exposure time of the current pixel G _{13 is} a short exposure. Here, for the current pixel G ₁₃ belonging to the short exposure, the up sampling circuit 102 adjusts the color values of the long exposure pixels such that the color values of the long exposure pixels correspond to the exposure time of the current pixel G ₁₃ . The step of the up sampling circuit 102 adjusting the color value of the long exposure pixel to correspond to the exposure time of the current pixel G ₁₃ may also be referred to as upsampling. Specifically, when the up sampling is performed, the up sampling circuit 102 divides the color value of each long exposure pixel by a gain (this gain is, for example, the ratio of the long exposure time to the short exposure time described above). On the other hand, the upsampling circuit 102 maintains the color values of the short exposure pixels unchanged. In this way, the upsampling circuit 102 can obtain the short exposure block W _S , wherein the pixels G′ ₁ , B′ ₂ , G′ ₃ , B′ ₄ , G′ ₅ , R′ ₆ , G′ ₇ , R′ _{8 , G '9, R' 10} , G '21, B' 22, G '23, B' 24 and pixel G 'color value of ₂₅ corresponds to a short exposure time.

It should be noted that, although the present embodiment is exemplified by the current pixel G _{13 which} is a short exposure, the present embodiment does not limit the exposure time of the current pixel G ₁₃ to be a short exposure. In another embodiment, the exposure time of the current pixel G ₁₃ may also be a long exposure. In this alternate embodiment, the upsampling circuit 102 adjusts the color values of the short exposure pixels in the operational block to correspond to long exposure times. For example, the upsampling circuit 102 multiplies the color values of each of the short exposure pixels by a gain. This gain is calculated based on the long exposure time and the short exposure time (for example, the ratio of the long exposure time to the short exposure time).

Upsampling circuit 102 also determines whether the operation of the pixel block W ₁ overexposure to generate a determination result. Specifically, the sampling circuit 102 will rise to obtain _an operation in addition to the W block of pixels of a pixel G ₁₃ (also known as test pixel). If a test pixel has a long exposure, the upsampling circuit 102 determines whether the color value of the test pixel is greater than the first threshold. If the color value of the test pixel is greater than the first threshold, the upsampling circuit 102 determines that the test pixel is overexposed, and the upsampling circuit 102 will mark the test pixel as overexposed in the determination result 310. If a test pixel has a short exposure, the upsampling circuit 102 determines whether the product of the color value of the test pixel multiplied by the gain is greater than the second threshold. If the product is greater than the second threshold, the upsampling circuit 102 determines that the test pixel is overexposed, and the upsampling circuit 102 will mark the test pixel as overexposed in the decision result 310. However, the present invention does not limit the first threshold value and the second threshold value. Here, the judgment result 310 is represented by a matrix (or a mask), wherein a value of "1" indicates that the corresponding test pixel has overexposure, and a value of "0" indicates that the corresponding test pixel has not been overexposed. For example, pixel G ₇ and pixel G ₉ are overexposed. It is to be noted that the judgment result 310 also indicates whether or not one pixel in the short exposure block Ws is overexposed. That is, the pixel G' ₇ and the pixel G' ₉ are marked as overexposed.

In another embodiment, the exposure time of the pixel G13 is currently a long exposure, and the up sampling circuit 102 also produces a determination result. The method for generating the judgment result is the same as the method for generating the judgment result 310 described above, and therefore will not be described again.

Referring to FIG. 1 and FIG. 3, the detection circuit 104 is coupled to the connected sampling circuit 102 liters, wherein the current detection circuit 104 for determining whether a pixel is a bad pixel G _13. For the current pixel G ₁₃ belonging to the short exposure, the detecting circuit 104 determines whether the current pixel G ₁₃ is a bad pixel according to the short exposure pixel in the short exposure block W _S , the adjusted long exposure pixel and the determination result 310 . . Moreover, the detection circuit 104 performs the determination based on the pixels (also referred to as the third pixels) having the same channel as the current pixel G ₁₃ . In this embodiment, since the pixel G ₁₃ has the channel G, the third pixel is the pixels G′ ₁ , G′ ₃ , G′ ₅ , G′ ₇ , G′ ₉ , G ₁₁ , G ₁₃ , G ₁₅ , G ₁₇ , G ₁₉ , G' ₂₁ , G' ₂₃ and pixel G' ₂₅ . The detecting circuit 104 determines whether each third pixel is overexposed according to the determination result 310. For each third pixel, if the corresponding third pixel in the determination result 310 is overexposed, the detection circuit 104 updates a count value (for example, plus 1).

On the other hand, for each third pixel, if the corresponding third pixel in the determination result 310 is not overexposed, the detecting circuit 104 may use the color value of the corresponding third pixel and the color value of the current pixel G ₁₃ . The gap between them is to update the above count value. For example, the detection circuit 104 calculates a third pixel (e.g., pixel G ₁₎ were the color value difference between the current pixel color values G _13. If the difference is greater than a fourth threshold, the detection circuit 104 updates the above count value (eg, plus one). If the difference is less than a fifth threshold, the detection circuit 104 also updates the count value (eg, plus one). If the difference between the color values of the pixel G ₁ and the pixel G ₁₃ is between the fourth critical value and the fifth critical value, the detecting circuit 104 maintains the count value unchanged. The detecting circuit 104 can also determine the fourth threshold value and the fifth threshold value according to the color value of the current pixel G ₁₃ . For example, when the color value of the current pixel G ₁₃ is larger, the fourth critical value will be larger and the fifth critical value will be smaller. However, the present invention does not limit the fourth critical value and the fifth critical value.

When the above count value is larger, it indicates that the pixel P _{13 is} more likely to be a bad pixel. The detecting circuit 104 determines whether the above count value is greater than a third critical value. If the count value is not greater than the third threshold, the detecting circuit 104 determines that the current pixel G _{13 is} not a bad pixel, and outputs the current pixel G ₁₃ . If the above count value is greater than the third critical value, it indicates that the current pixel G ₁₃ is a bad pixel. In particular, the detecting circuit 104 further determines whether one of the third pixels is overexposed, thereby repairing the current pixel G ₁₃ in a different manner. Specifically, if any one of the third pixels is overexposed, the detecting circuit 104 determines that the current pixel G ₁₃ is a bad pixel, and the repair circuit 106 has the same channel and the same exposure according to the current pixel G ₁₃ . The third pixel of time corrects the current pixel G ₁₃ . For example, suppose that the current pixel G ₁₃ is judged to be a bad pixel, and there is a third pixel (ie, the pixel G' ₇ or the pixel G' ₉ ) is overexposed, so the repair circuit 106 according to the short exposure The three pixels, that is, the pixels G ₁₁ , G ₁₅ , G ₁₇ and the pixel G ₁₉ , correct the current pixel G ₁₃ . In an embodiment, the repair circuit 106 performs the above-described correction of the current pixel G ₁₃ according to, for example, a weighted average algorithm. Specifically, the repair circuit 106 performs a weighted average algorithm based on the color values of the third pixels G ₁₁ , G ₁₅ , G _{17 ,} and the pixels G ₁₉ having a short exposure time, and obtains results based on performing the weighted average algorithm. To correct the color value of the current pixel G ₁₃ . Accordingly, in consideration of overexposure, repair circuit 106 does not use overexposed pixel G _'7 and the pixel G' ₉ to repair the current pixel G _13.

Conversely, if the current pixel G ₁₃ is a bad pixel and all of the third pixels are not overexposed, the repair circuit 106 corrects the current pixel G ₁₃ according to all the pixels in the short exposure block Ws. However, the repair circuit 106 can correct the current pixel G ₁₃ according to any algorithm, and the present invention is not limited thereto.

It should be noted that, in another embodiment, the detecting circuit 104 also determines whether it is a bad pixel for the current pixel belonging to the long exposure according to a similar method. That is to say, the detection circuit 104 updates the count value according to the third pixel belonging to the long exposure and having the same channel as the current pixel. The detection circuit 104 also updates the count value based on whether the third pixel is overexposed. And the repair circuit 106 will follow The way to correct the current pixels that belong to long exposures. That is, if the current pixel belonging to the long exposure is determined by the detecting circuit 104 to be a bad pixel and one pixel is overexposed, the repair circuit 106 will have a third channel having the same channel as the current pixel and having a long exposure time. Pixels to correct this current pixel.

It is noted that the upsampling circuit 102, the detecting circuit 104, and the repairing circuit 106 of the present embodiment can be implemented as a hardware circuit composed of one or several logic gates. Alternatively, in another embodiment of the present invention, the functions of the upsampling circuit 102, the detecting circuit 104, and the repairing circuit 106 can be implemented by computer code. The computer code is stored in a storage unit of the electronic device and executed by a processing unit. The processing unit is a hardware (eg, a chipset, a processor, etc.) having computing power for controlling the overall operation of the electronic device 300. The processing unit is, for example, a central processing unit (CPU), or other programmable microprocessor (Microprocessor), a digital signal processor (DSP), a programmable controller, and a special application product. Application Specific Integrated Circuits (ASICs), Programmable Logic Devices (PLDs), or the like. In addition, the above storage unit may be an in-line storage unit or an external storage unit. The embedded storage unit can be a random access memory (RAM), a read-only memory (ROM), a flash memory, a magnetic disk storage device (Magnetic disk storage). Device) and so on. The external storage unit can be a compact flash (CF) memory card, a secure digital (SD) memory card, and a micro-safe digital (Micro SD) memory. Card, Memory Stick (MS), etc. In this embodiment, the storage unit may store one or more codes, materials, and the like for performing the load prediction method.

FIG. 4 is a flowchart of a bad pixel processing method according to an embodiment of the invention. Referring to FIG. 4, in step S401, a plurality of pixels are obtained, wherein a plurality of first pixels in the pixels have a first exposure time, and a plurality of second pixels in the pixels have a second exposure time, first The exposure time is different from the second exposure time, and the current pixel in the pixel belongs to the first exposure time. In step S403, the color values of the respective second pixels are adjusted to correspond to the first exposure time. In step S405, it is determined whether the pixel is overexposed to generate a determination result. In step S407, it is determined whether the current pixel is a bad pixel according to the first pixel, the adjusted second pixel, and the determination result. In step S409, if the current pixel is a bad pixel, the current pixel is corrected. Thereby, the present invention can facilitate the bad pixel detection and repair processing of high dynamic range images to produce high quality high dynamic range images. It should be noted that, in an embodiment, the first exposure time and the first pixel are, for example, the long exposure time and the long exposure pixel described above. However, the present invention is not limited thereto, and the first exposure time and the first pixel may also be a short exposure time and a short exposure pixel. The steps in Fig. 4 have been described in detail above and will not be described again here. It should be noted that the steps in FIG. 4 can be implemented as a plurality of codes or circuits; in addition, the method of FIG. 4 can be used in combination with the above embodiments, or can be used alone, and the present invention is not limited thereto.

In order to clarify the above-described bad pixel processing method, an embodiment will be described below. FIG. 5 is a flowchart of a bad pixel processing method according to an embodiment of the invention.

Referring to FIG. 5, in step S501, a plurality of pixels in the operation block are obtained according to the current pixel to be processed. A plurality of first pixels of the pixels have a first exposure time, and a plurality of second pixels of the pixels have a second exposure time. The first exposure time is different from the second exposure time, and the current pixel in the pixel belongs to the first exposure time.

In step S503, the color values of the respective second pixels are adjusted to correspond to the first exposure time, and it is determined whether the pixels are overexposed to generate a determination result.

In step S505, it is determined according to the determination result whether each third pixel in the operation block is overexposed, wherein the third pixel has the same channel as the current pixel. For each third pixel, if the corresponding third pixel is overexposed, a count value is updated. If there is a third pixel that is not overexposed, the difference between the color value of the third pixel and the color value of the current pixel is used to update the count value. Finally, it is determined whether the count value is greater than the third critical value.

If the count value is not greater than the third threshold, as shown in step S507, it is determined that the current pixel is not a bad pixel, and the current pixel is output.

However, if the count value is greater than the third threshold, then as shown in step S509, it is determined whether one of the third pixels is overexposed.

Here, if none of the third pixels is overexposed, as shown in step S511, the current pixel is corrected by a single image processing method based on a single image, and the corrected current pixel is output. Step S511 can use any algorithm to correct the current pixel, and the present invention is not limited thereto.

However, if any of the third pixels is overexposed, Step S513, determining that the current pixel is a bad pixel, and correcting the current pixel according to the third pixel having the first exposure time.

It should be noted that, in an embodiment, the first exposure time and the first pixel are, for example, the long exposure time and the long exposure pixel described above. However, the present invention is not limited thereto, and the first exposure time and the first pixel may also be a short exposure time and a short exposure pixel. The steps in Fig. 5 have been described in detail above, and will not be described again here.

In summary, in the bad pixel processing method and the image processing apparatus proposed in the embodiments of the present invention, the pixels around the current pixel are upsampled to obtain pixel information consistent with the exposure time of the current pixel, and the Mark the pixels with overexposure to get a judgment result, and finally use the above-mentioned upsampled pixel information and the judgment result to perform bad pixel detection and repair. Thereby, the present invention can facilitate bad pixel detection and repair processing in images with different exposure times.

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.

S401, S403, S405, S407, S409‧‧‧ steps of the bad pixel processing method

Claims (16)

A method for processing a bad pixel for an image processing apparatus includes: acquiring a plurality of pixels, wherein a plurality of first pixels of the pixels have a first exposure time, and a plurality of second pixels of the pixels have a a second exposure time, the first exposure time is different from the second exposure time, and one of the pixels belongs to the first exposure time; adjusting a color value of each of the second pixels to correspond to the first Determining whether the pixels are overexposed to generate a determination result; determining whether the current pixel is a bad pixel according to the first pixels, the adjusted second pixels, and the determination result; The current pixel is the bad pixel, and the current pixel is corrected. The bad pixel processing method of claim 1, wherein the first exposure time is one of a short exposure and a long exposure, and the second exposure time is one of the short exposure and the long exposure. In addition, the step of adjusting the color value of each of the second pixels to correspond to the first exposure time comprises: if the second exposure time is the short exposure, the color value of each of the second pixels Multiplying a gain, wherein the gain is calculated according to the first exposure time and the second exposure time; and if the second exposure time is the long exposure, dividing the color value of each of the second pixels by The gain. The bad pixel processing method of claim 2, wherein the step of determining whether the pixels are overexposed to generate the determination result comprises: Obtaining a test pixel of the pixels; if the test pixel has the long exposure, determining whether a color value of the test pixel is greater than a first threshold, and if the color value of the test pixel is greater than the first threshold Determining whether the test pixel is overexposed; if the test pixel has the short exposure, determining whether the product of the color of the test pixel multiplied by the gain is greater than a second threshold, and if the product is greater than the second a threshold value, determining that the test pixel is overexposed; and marking whether the test pixel is overexposed in the determination result. The method of processing a bad pixel according to claim 1, wherein the plurality of third pixels of the pixels have the same channel as the current pixel, and the step of determining whether the current pixel is the bad pixel comprises: Determining, according to the determination result, whether each of the third pixels is overexposed; for each of the third pixels, if the corresponding third pixel is overexposed, updating a count value; for each of the third pixels a pixel, if the corresponding third pixel is not overexposed, updating the count value according to a gap between a color value of the corresponding third pixel and a color value of the current pixel; determining whether the count value is And greater than a third threshold; and if the count value is not greater than the third threshold, determining that the current pixel is not the bad pixel. The bad pixel processing method of claim 4, wherein the color value according to the corresponding third pixel and the color value of the current pixel are The step of updating the count value includes: determining a fourth threshold value and a fifth threshold value according to the color value of the current pixel; if the gap is greater than the fourth threshold value, updating the count value; if the gap is If the gap is less than the fifth threshold, the count value is updated; and if the gap is between the fourth threshold and the fifth threshold, the count value is maintained unchanged. The method of processing the bad pixel according to the fourth aspect of the invention, wherein the determining whether the current pixel is the bad pixel further comprises: if the count value is greater than the third threshold, determining the third pixel Whether it is overexposure; and if any of the third pixels is overexposed, determining that the current pixel is the bad pixel. The bad pixel processing method of claim 6, wherein the step of modifying the current pixel comprises: correcting the current pixel according to the third pixels having the first exposure time. The bad pixel processing method of claim 7, wherein the step of correcting the current pixel according to the third pixels having the first exposure time is performed according to a weighted average algorithm. An image processing apparatus includes: a one-liter sampling circuit for acquiring a plurality of pixels, wherein a plurality of the pixels The first pixel has a first exposure time, and the plurality of second pixels of the pixels have a second exposure time, the first exposure time is different from the second exposure time, and one of the pixels belongs to the current pixel The first exposure time, wherein the up sampling circuit is configured to adjust a color value of each of the second pixels to correspond to the first exposure time, and determine whether the pixels are overexposed to generate a determination result; The circuit is coupled to the upsampling circuit for determining whether the current pixel is a bad pixel according to the first pixels, the adjusted second pixels, and the determination result; and a repair circuit coupled to the The detecting circuit is configured to correct the current pixel if the current pixel is the bad pixel. The image processing device of claim 9, wherein the first exposure time is one of a short exposure and a long exposure, and the second exposure time is one of the short exposure and the long exposure. The operation of adjusting the color value of each of the second pixels to correspond to the first exposure time includes: if the second exposure time is the short exposure, the up sampling circuit will each The color value of the two pixels is multiplied by a gain, wherein the gain is calculated according to the first exposure time and the second exposure time; and if the second exposure time is the long exposure, the up sampling circuit will each The color value of the second pixels is divided by the gain. The image processing device of claim 10, wherein the upsampling circuit determines whether the pixels are overexposed to generate the determination result comprises: The up sampling circuit obtains one of the pixels; if the test pixel has the long exposure, the up sampling circuit determines whether a color value of the test pixel is greater than a first threshold, and if the test pixel The color value is greater than the first threshold, the up sampling circuit determines that the test pixel is overexposed; if the test pixel has the short exposure, the up sampling circuit determines a product of the color value of the test pixel multiplied by the gain Whether it is greater than a second threshold value, and if the product is greater than the second threshold value, the up sampling circuit determines that the test pixel is overexposed; and the up sampling circuit marks in the determination result whether the test pixel is overexposed. The image processing device of claim 9, wherein the plurality of third pixels of the pixels have the same channel as the current pixel, and the detecting circuit determines whether the current pixel is the bad pixel. The detecting circuit determines whether each of the third pixels is overexposed according to the determination result; and for each of the third pixels, if the corresponding third pixel is overexposed, the detecting circuit updates the meter a value; for each of the third pixels, if the corresponding third pixel is not overexposed, the detecting circuit is configured according to a color value of the corresponding third pixel and a color value of the current pixel Updating the count value by a gap; the detecting circuit determines whether the count value is greater than a third threshold; and if the count value is not greater than the third threshold, the detecting circuit determines the current image It is not the bad pixel. The image processing device of claim 12, wherein the detecting circuit updates the count value according to the difference between the color value of the corresponding third pixel and the color value of the current pixel. The detecting circuit determines a fourth critical value and a fifth critical value according to the color value of the current pixel; if the difference is greater than the fourth critical value, the detecting circuit updates the counting value; The detection circuit updates the count value when the difference is less than the fifth threshold value; and if the difference is between the fourth threshold value and the fifth threshold value, the detecting circuit maintains the count value unchanged. The image processing device of claim 12, wherein if the count value is greater than the third threshold, the detecting circuit further determines whether one of the third pixels is overexposed, if the Any one of the three pixels is overexposed, and the detecting circuit determines that the current pixel is the bad pixel. The image processing device of claim 14, wherein the repairing circuit corrects the current pixel according to the third pixels having the first exposure time. The image processing device of claim 15, wherein the repairing circuit performs the operation of correcting the current pixel according to the third pixels having the first exposure time according to a weighted average algorithm.