TWI513310B - Device and method for expanding dynamic range of camera - Google Patents

Description

Camera dynamic range expansion device and method

The present invention relates to a camera dynamic range expansion device and method, and more particularly to a camera dynamic range expansion device and method for facilitating generation of a high dynamic range image by using an algorithm designed by a system device.

Press, digital cameras are widely used in life, such as: driving recorders, monitors and digital cameras...etc. When shooting, the image captured by the lens is projected onto the photosensitive element. Due to the exposure design of the existing camera photosensitive element, if the source of the light source is extremely extreme, the visible range is often narrowed, that is, the entire screen has only the bright part. Presented, the dark part is black, or the whole picture is only in the dark part, and the bright part is white, and the brightness is too dark.

When the photographic picture is only partially illuminated, and the dark part is black, the sensing component of the camera is exposed to the brighter area of the scene. Because of the high brightness, the camera sensing component will select a shorter exposure time to prevent it. Excessive exposure, in order to accommodate the short exposure time of the bright area, the dark area is therefore underexposed, and appears a black, so that the brightness is too dark, the image can not be seen clearly, and vice versa, the entire photographic picture is only in the dark part, The bright part is white, because the sensing component of the camera is exposed to the darker area of the scene, because the brightness is not Foot, the sensing element of the camera will choose a longer exposure time to make the exposure sufficient, and in order to accommodate the long exposure time of the dark area, the bright area is overexposed and appears white, resulting in an image of the bright area It cannot be clearly displayed, this is the Low Dynamic Range (LDR) image.

In order to overcome this low dynamic range image problem, existing photographers often take different exposure value settings during shooting to capture the same image, and after several images with different sensitivity levels, take out each detail. The part is synthesized into the same piece. Since the correct exposure time is taken for each part of the photographic scene to collect the image information, the true darkness effect can be obtained. However, this method requires storing multiple images and causing the camera. The image storage space is wasted, and when the synthesis process is performed, it is easy to be technically good by the photographer, and there is a difference in quality, and it takes a considerable amount of time to perform the synthesis process, and does not conform to the efficiency.

The reason is that the present inventors have in view of the existing ways of overcoming the low dynamic range image, and the above-mentioned many disadvantages are supported by the manufacturing and design experience and knowledge of the related fields for many years, and through various ingenuity, The dynamic range expansion of the camera has been developed and improved, and the present invention has been developed.

The present invention relates to a camera dynamic range expansion apparatus and method, and its main object is to provide a camera dynamic range expansion apparatus and method for facilitating generation of a high dynamic range image by using an algorithm designed by a system apparatus.

In order to achieve the above-mentioned implementation purposes, the inventors have developed the following camera dynamic range expansion device, which mainly includes a color space conversion module, a parameter statistics circuit, a type judgment circuit, a dark portion reinforcement circuit, an inversion processing module, a hybrid circuit, and The bright portion lifting circuit couples the color space conversion module with the parameter statistical circuit, and couples the parameter statistical circuit with the type determining circuit, and enables the The type judging circuit is coupled to the dark portion reinforcing circuit and the hybrid circuit, and the dark portion reinforcing circuit is coupled to the hybrid circuit, and the color space conversion module is coupled to the inversion processing module, and The inverting processing module is coupled to the hybrid circuit, and the hybrid circuit is coupled to the bright portion boosting circuit.

The camera dynamic range expansion device as described above, wherein the camera dynamic range expansion device further comprises an automatic gain control circuit coupled to the bright portion boost circuit.

The camera dynamic range expansion device as described above, wherein the inversion processing module comprises a phase-coupling filling circuit, a filtering circuit and an inversion processing circuit, so that the color space conversion module is filled with The circuit is coupled to the circuit, and the inverting processing circuit is coupled to the hybrid circuit.

The invention further proposes a camera dynamic range expansion method, the implementation steps are as follows: A. Image color space conversion: the image taken by the camera is input into a color space conversion module to red, green, The color space of the blue primary color is converted into the color space of brightness and chromatic aberration; B. Image type classification: following the brightness value data of the original image is sent to a parameter statistical circuit for histogram statistical analysis, and then a type judgment circuit The original image is classified into one of too dark type, over bright type or extreme type; C. Image dark part reinforcement processing: according to the type judgment circuit, the type of judgment is different, and the image of the over dark type and the extreme type is conveyed Reinforce the circuit to a dark part to perform the dark part reinforcement operation to obtain the dark part reinforcement image, and then input a hybrid circuit. If it is a super bright image, input the hybrid circuit directly; D. Image inversion processing: another color Space conversion module conversion The original image data is simultaneously sent to an inversion processing module for performing inversion processing through the inversion processing module to obtain an inverted image to input the hybrid circuit; E. Image mixing processing: continued by The hybrid circuit mixes the input inverted image and the dark portion enhanced image to obtain a mixed image; F. Image bright portion lifting process: after a bright portion lifting circuit performs the bright image lifting operation , get a high dynamic range image.

The camera dynamic range expansion method as described above, wherein the over-dark, over-bright, and extreme types in the image type classification step are based on two parameter values of Hi and Lo, and the Hi and Lo two parameter values are used. The formula is as follows:

Where h(i) represents the statistical pixel quantity of the histogram gray scale i, and Th1~Th4 are four custom thresholds, respectively falling in four positions of the histogram gray scale from small to large, when Lo > 2 Hi , the system is too dark, and when Hi > 2 Lo , it is too bright, another or When it is extreme, the P _total is the total number of pixels of the picture.

The camera dynamic range expansion method as described above, wherein the type judging circuit further comprises a first multiplexer, a second multiplexer and a third multiplexer coupled to each other, and the type judging circuit judges Lo >2 In Hi , the system shifts Hi by 1 bit to get 2Hi, and then compares it with Lo. If Lo > 2 Hi is met, the first multiplexer selects 1, and its serial number (type) is output 00 represents too dark. Type, if it does not meet Lo > 2 Hi , the first multiplexer selects 0. At this time, the output is the result of the second multiplexer, and the second multiplexer determines whether it meets the over-bright condition Hi > 2 Lo . Control, if it is met, the second multiplexer selects 1, the serial number output is 01 for over-bright type. If Hi > 2 Lo is not met, the second multiplexer selects 0. At this time, its output is the third multiplex. As a result, the third multiplexer determines whether it meets the extreme conditions. or To control, if any of the above two conditions are met, the third multiplexer selects 1. At this time, the serial number output 10 represents the extreme type. If both conditions are not met, the serial number output 11 represents the normal type.

The camera dynamic range expansion method as described above, wherein the inverse image and the dark portion enhancement image are multiplied in the image mixing processing step, and the multiplication operation formula is as follows: y _mix ( i , j )= y _dark ( i , j )× y _inv ( i , j )× k

The y _mix is the output value after the hybrid operation, y _dark is the dark portion to reinforce the image value, y _inv is the inverted image value, and k is the key parameter for outputting the high dynamic range image.

The camera dynamic range expansion method as described above, wherein the image type classification step further comprises an operation of a nonlinear curve equation, and the formula of the nonlinear curve equation is as follows:

Where y _inv is the inverse value and y _lpf is the input value of the original image data.

The camera dynamic range expansion method as described above, wherein the image type classification step further comprises an operation of compensating the equation, and the formula of the compensation equation is as follows:

Where y _dark is the dark portion to reinforce the image output value, y _in is the input image grayscale value, and Th _dark is the custom threshold.

The camera dynamic range expansion method as described above, wherein the bright portion lifting operation formula of the image bright portion lifting processing step is as follows:

Where y _mix is the mixed image value, y _dark is the dark portion to reinforce the image value, Th _b is the threshold value, and y _HDR is the high dynamic range image value.

The camera dynamic range expansion method as described above, wherein the dark portion reinforcing circuit further comprises a multiplexer coupled to the multiplexer, the register and the subtractor, and the multiplexer of the dark portion reinforcing circuit receives the output from the type determining circuit When one of the dark type numbers is over, the multiplexer selects 128 as the threshold value output. When the multiplexer of the dark portion reinforcing circuit receives an extreme type number output by the type determining circuit, the multiplexer selects 50 as the valve. Value output, the output value is stored in the register, and then the buffer sends the threshold to the subtractor and subtracts the image brightness value data input by the type judgment circuit, and subtracts the result right After shifting 4 bits, the image of the input brightness value is added to obtain the result of the reinforcement operation.

The camera dynamic range expansion method as described above, wherein the bright portion lifting circuit comprises a phase coupled multiplexer, a comparator, a subtractor, a multiplier and two adders, so that the subtractor performs one Dark portion reinforcement value input by the dark portion reinforcement circuit [dark_enh] minus 50 action, the multiplier multiplies the result of the subtractor by itself to obtain the dark squared square value [dark_enh_mul], and then uses the right shift 8 bit and the right shift 10 bit to add the multiplication The value of 0.005 is then added to the output value [out_mix] of the mixed image, and the rear multiplexer [MUX] is selected to output.

Thereby, when the camera lens captures the image and inputs the color space conversion module of the present invention, the parameter statistics circuit, the type judgment circuit, the dark portion reinforcement circuit, the inversion processing module, the hybrid circuit, the bright portion lifting circuit, and the automatic The operation of the gain control circuit can instantly output images with high dynamic range (HDR), without waiting for the time for manual processing, which is in line with practical efficiency, and can also avoid relying on manual synthesis processing because of the technically good processing personnel.窳, the quality difference is generated, and the invention only needs a single image data to generate a high dynamic range image, and can avoid storing a plurality of images, resulting in improper storage space and the like.

(1) ‧‧‧Color Space Conversion Module

(2) ‧‧‧Parameter statistics circuit

(3) ‧‧‧ type judgment circuit

(4) ‧‧‧Dark part reinforcement circuit

(5) ‧‧‧ reverse processing module

(51)‧‧‧ Filling the circuit

(52)‧‧‧Filter circuit

(53)‧‧‧Inverting Processing Circuit

(6) ‧‧‧Hybrid circuits

(7) ‧ ‧ bright boost circuit

(8)‧‧‧Automatic gain control circuit

First figure: the architecture diagram of the present invention

Second figure: flow chart of the present invention

The third figure: the Lo value statistical circuit diagram of the present invention

Fourth figure: the type judgment circuit diagram of the present invention

Figure 5: Reinforcement circuit diagram of the dark portion of the present invention

Figure 6: Bright circuit diagram of the bright part of the present invention

In order to make the technical means of the present invention and the effects thereof can be more completely and clearly disclosed, the details are as follows, please refer to the disclosure. Figure and figure number: First, please refer to the first figure, which is the camera dynamic range expansion device of the present invention, which mainly consists of a color space conversion module [RGB to YCbCr Module] (1), parameter statistics circuit [Parameter Statistics] Circuit] (2), Type Judgment Circuit (3), Dark Enhancement Circuit (4), Inverse Module (5), Mix Circuit (Mix Circuit) 6), a Bright Enhancement Circuit (7) and an Automatic Gain Control Circuit (8); wherein: the color space conversion module (1) is associated with the parameter statistics circuit ( 2) coupled, and the parameter statistical circuit (2) is coupled to the type determining circuit (3), and the type determining circuit (3) and the dark portion reinforcing circuit (4) and the hybrid circuit ( 6) coupling, the dark portion reinforcing circuit (4) is coupled to the hybrid circuit (6), and the inverting processing module (5) includes a phase-parameterly coupled filling circuit (51) and a filter circuit. (52) and an inverse processing circuit (Inverse Processing Circuit) (53), The filling circuit (51) is a 2×2 maximum filling circuit [2×2 Maximum Fill Circuit], and the filtering circuit (52) is a 3×3 low-pass filter circuit (3×3 Lowpass Filter), The filling circuit (51) of the inverting processing module (5) is coupled to the color space conversion module (1), and the inverting processing circuit (53) of the inverting processing module (5) is The hybrid circuit (6) is coupled to the hybrid circuit (7) and coupled to the bright portion boosting circuit (7), and the bright portion boosting circuit (7) is coupled to the automatic gain control circuit (8) .

Accordingly, please refer to the second figure, the implementation steps of the camera dynamic range expansion device of the present invention include the following: A. Image color space conversion: the original image taken by the camera lens [ y _lpf After inputting the color space conversion module (1) of the present invention, the color space conversion module (1) of the present invention converts the color spaces of the three primary colors [RGB] of red, green and blue into the color of the brightness and the color difference [YCbCr]. space.

B. Image type classification: the original image [ y _lpf ] brightness value [Y] data is sent to the parameter statistics circuit (2) for histogram statistical analysis [Histogram Analysis], and then by the type judgment circuit (3) The original image [ y _lpf ] is classified into one of three types: Dark Type, Bright Type, or Extreme Type, and the over-dark, over-bright, and The extreme type is based on the Hi and Lo parameters. The Hi and Lo parameters are as follows:

Where h(i) represents the statistical pixel quantity of the histogram gray scale i, and Th1~Th4 are four custom thresholds, respectively falling in four positions of the histogram gray scale from small to large. The obtained Hi and Lo values can provide the type judgment circuit (3) as the classification of the image type, and the over dark type indicates that the entire image is dark, so the Lo value is much higher than the Hi value, when Lo > 2 Hi , The type judging circuit (3) judges to be too dark, whereas the over-bright type indicates that the entire image is bright. Therefore, the Hi value is much higher than the Lo value. When Hi > 2 Lo , the type judging circuit (3) Judging as too bright, the other extremes represent a lot of bright parts of the entire image, too many parts of the dark, so the Hi value and Lo value will be quite high, but the difference between them is not more than twice the other, so when or The type judging circuit (3) judges to be an extreme type, wherein the P _total is the total number of pixels of the picture.

C. Image dark portion reinforcement processing: According to the type judgment circuit (3), the type of judgment is different, and the image of the over dark type and the extreme type is sent to the dark portion reinforcing circuit (4) for dark enhancement (Dark Enhancement) action, due to extreme The darkness of the type is less than the over-dark type. Therefore, when the dark portion is compensated, the dark portion compensation processing is performed with the parameter smaller than the over-dark type to obtain the dark portion reinforcing image [ y _dark ], and then input to the hybrid circuit (6). The over-bright image is sent directly to the hybrid circuit (6).

D. Image inversion processing: the original image [ y _lpf ] data converted by the color space conversion module (1) is simultaneously sent to the inversion processing module (5) to be processed by the inversion processing module ( 5) the fill circuit (51), a filter circuit (52) processing circuit and an inverter (53) to give the inverted image] [y _inv delivery to the mixing circuit (6).

E. Image mixing processing: The mixed image ( y _inv ) and the dark portion reinforcing image ( y _dark ) are mixed by the mixing circuit (6) to obtain the corrected mixed image [ y _mix ] The mixing process of the inverse image [ y _inv ] and the dark portion enhancement image [ y _dark ] is to multiply the two images by the following equation: y _mix ( i , j )= y _dark ( i , j )× y _inv ( i , j )× k

Where y _mix is the output value after the blending operation, y _dark is the dark portion to reinforce the image value, y _inv is the inverse image value, and k is the key parameter for outputting the high dynamic range image, the larger the k value is available The higher the dynamic range (High Dynamic Range, HDR), the more the dark portion is raised, the more the bright portion is also reduced. The more the output is, the inverse image [ y _inv ] and the dark portion image ( y _dark ) are used. After multiplying, the mixed image [ y _mix ] will be brightened in the dark and the bright spot will be reduced.

F. Image brightening processing: In order to prevent the mixing process from being weakened, the original brightness is weakened to be darker than the original brightness, and the present invention is the mixed image after the mixing process [ y _mix 〕 is sent to the bright portion lifting circuit (7) to perform the Bright Enhancement action by the bright portion lifting circuit (7) to prevent excessive darkening of the original bright portion, and the calculation formula is as follows:

Where y _{mix is} the main processing object of the formula, that is, the mixed image value after the blending process, y _dark is the dark portion reinforcing image value, Th _b is the threshold value, and y _HDR is the high dynamic range graph after the bright portion is raised. like value to the calculation formula for mainly reinforcing the dark portion of the image] [y _dark gray level value greater than Th _b adjust the mixed image portion [of] y _mix, to obtain a high dynamic range image] [y _HDR The system will then send it to the automatic gain control circuit (8) for brightness contrast adjustment and output, and then the fine-tuned high dynamic range image output can be obtained.

In order to solve the problem that the parameter statistical circuit (2) of the present invention is excessively improved in the middle portion of the gray scale value in the histogram, the present invention utilizes a nonlinear curve equation, and the formula is as follows:

Where y _inv is the inverse value and y _lpf is the original image input value. In the nonlinear curve equation, the lower the input value [ y _lpf ], the higher the output value [ y _inv ], and the lower gray scale. , can generate a higher inverse gray scale value to multiply to increase the brightness, if the original image input value [ y _lpf ] is 0, although the maximum inverse value of 255 will be obtained, because the present invention is in the hybrid circuit (6) the mixed image and the original image line by the inverter y _lpf [image] [image reinforcing derived by multiplying] y _dark dark dark portion of the reinforcing portion, when the original image is input [0] y _lpf, resulting in system The multiplied product is 0, and the brightness cannot increase the missing. Therefore, in order to prevent the above, the present invention uses the following formula to perform the compensation process:

Where y _dark is the dark part to reinforce the image output value, y _in is the input image gray scale value, the threshold value in the formula [ Th _dark ] is the amplitude that can change the brightness, and the higher the threshold value ( Th _dark ) makes the low gray The higher the gain of the gradation, the wider the processing range and the wider the influence range, and the lower the threshold ( Th _dark ) can converge the range of influence, and the increase for the low gray level is smaller, if the input is An over-dark image indicates that the image is mostly low-gray. In this case, a larger threshold ( Th _dark ) is used to increase the gain and processing range. The present invention is a valve for over-dark images. The value of [ Th _dark ] is set to 128. If the input is an extreme image, the low gray level in the extreme image is less than that of the dark image. If the value is the same as the over dark image, the threshold is used. [ Th _dark ] will increase the darkness of the image too much, so the threshold value of the extreme image ( Th _dark ) should be lower than that of the dark image.

The condition of the present invention is judged according to two parameters Lo and Hi, and the Lo value is a histogram whose gray scale is less than 50, wherein the gray scale is less than 15 and the double is added. Since the Lo and Hi values are statistical images of the entire image [ After the frame], the parameters can be obtained. According to the real-time output requirement on the hardware, the current image [frame] data cannot be counted and then processed. Therefore, the present invention uses the previous image [frame] statistics. Make the Lo value of the current image [frame]. Please refer to the third figure, the invention first determines whether the input value is less than 15 in the circuit execution, and if so, performs a +2 action, otherwise it continues to determine whether the input value is less than 50, and if yes, then Perform a +1 action and repeat the above actions until the vertical sync signal is triggered, that is, the end of an image [frame]. At this time, the accumulated value output is used as the Lo value for the next image [frame], and then the zero accumulator is used. Recalculate the next Lo value.

Please also refer to the fourth figure, the type judging circuit (3) of the present invention includes a first multiplexer, a second multiplexer and a third multiplexer coupled to each other, when the type judging circuit (3) The judgment of the over-dark condition is Lo > 2 Hi . After shifting Hi to the left by 1 bit [bit], 2Hi is obtained, and then compared with Lo. If Lo > 2 Hi is satisfied, the first multiplexer is used. MUX 1] select 1, whose serial number [type] is output 00 for over-dark type; if Lo > 2 Hi is not met, the first multiplexer [MUX 1] selects 0, at this time, the output is the second multiplexer [ MUX 2] As a result, the second multiplexer [MUX 2] is controlled by judging whether it is in compliance with the over-bright condition Hi > 2 Lo , and if it is met, the second multiplexer [MUX 2] selects 1, the serial number [type] Output 01 stands for over-bright type. If Hi > 2 Lo is not met, the second multiplexer [MUX 2] selects 0. At this time, the output is the result of the third multiplexer [MUX 3], and the third multiplexer [MUX] 3] is to judge whether it meets the extreme conditions or To control, if any of the above two conditions are met, the third multiplexer [MUX 3] selects 1. At this time, the serial number [type] output 10 represents the extreme type. If both conditions are not met, the serial number [type] is output. 11 represents the normal type.

Please also refer to the fifth figure. The dark portion reinforcing circuit (4) of the present invention comprises a multiplexer, a register and a subtractor coupled to each other, and the multiplexer of the dark portion reinforcing circuit (4) receives the multiplexer. The over-type number [type[1]] output by the type judgment circuit (3) is 0, and the extreme type number [type[1]] is 1, when the serial number [type[1]] is 0, the multiplexer [MUX] will select 128 as the threshold value ( Th _dark ) output; when the serial number [type[1]] is 1, the multiplexer [MUX] will select 50 as the threshold value ( Th _dark ) output, and the output value will be Stored in the scratchpad, then the scratchpad is sent to the subtractor and subtracted from the image brightness value data [data_Y] input by the type judging circuit, and the subtraction result is shifted right by 4 bits [bit] and The input image brightness value data [data_Y] is added to obtain the result of the reinforcement operation, and the right shift of 4 bits [bit] is intended to replace In the middle, multiply 0.005 action.

Referring to FIG. 6 together, the bright portion lifting circuit (7) of the present invention is mainly composed of a multiplexer, a comparator, a subtractor, a multiplier and two adders, and the first subtractor The part is continuously executed by the dark part reinforcement value (dark_enh) minus 50 input by the dark part reinforcement circuit (4), and the multiplier multiplies the result of the subtractor by itself to obtain the dark part reinforcement square value [dark_enh_mul] is [15:0], after which Then use the right shift 8-bit [bit] and the right shift 10-bit [bit] add action to achieve 0.0039 + 0.009 = 0.0048, approximately equal to multiply the value of 0.005, and then add this value to the output value of the mixed image. [out_mix], the part that is now the highlight promotion calculation Points, wait for the rear multiplexer [MUX] to select the output. The selection of the rear multiplexer [MUX] in the circuit is based on a comparator that determines whether the dark portion reinforcement value [dark_enh] is less than 50, and the comparator continuously performs the comparison of the dark portion reinforcement value [dark_enh] with 50, if the dark portion reinforcement value [dark_enh] Less than 50 means there is no need to do the bright part boost, the comparator output is 1, the multiplexer directly selects the output value of the mixed image [out_mix] as the output of the high dynamic range output value [out_HDR], and the dark portion reinforcement value [dark_enh] is larger than A value equal to 50 indicates that the position needs to be brightened. At this time, the comparator outputs 0, and the rear multiplexer [MUX] selects the high dynamic range output value (out_HDR) after the bright portion is boosted.

Thereby, the image is taken by the camera lens, and after inputting the color space conversion module (1) of the present invention, the parameter statistics circuit (2), the type judgment circuit (3), the dark portion reinforcement circuit (4), and the inversion processing are performed. Module (5), hybrid circuit (6), bright boost circuit (7) and automatic gain control circuit (8) can instantly output high dynamic range (HDR) images without waiting for human synthesis The processing time is in line with the practical efficiency, and it can also avoid the quality difference caused by the processing personnel in the human synthetic processing, and the invention can only generate a single image data in the implementation. The dynamic range image can avoid storing multiple images, causing improper storage space and other situations.

The foregoing embodiments or drawings are not intended to limit the implementation of the camera dynamic range expansion apparatus and method of the present invention. Any changes or modifications made by those skilled in the art should be regarded as not departing from the camera dynamics of the present invention. Range extension device and method patent category.

In summary, the embodiments of the present invention can achieve the expected functions, and the specific structures disclosed therein have not been seen in similar products, nor have they been disclosed before the application, and have fully complied with the requirements and requirements of the Patent Law. If you apply for an invention patent in accordance with the law, you are welcome to review it and grant a patent.

(1) ‧‧‧Color Space Conversion Module

(2) ‧‧‧Parameter statistics circuit

(3) ‧‧‧ type judgment circuit

(4) ‧‧‧Dark part reinforcement circuit

(5) ‧‧‧ reverse processing module

(51)‧‧‧ Filling the circuit

(52)‧‧‧Filter circuit

(53)‧‧‧Inverting Processing Circuit

(6) ‧‧‧Hybrid circuits

(7) ‧ ‧ bright boost circuit

(8)‧‧‧Automatic gain control circuit

Claims (13)

A camera dynamic range expansion device mainly comprises a color space conversion module, a parameter statistics circuit, a type judgment circuit, a dark portion reinforcement circuit, an inverse processing module, a hybrid circuit and a bright portion lifting circuit, so that the color space conversion mode The group is coupled to the parameter statistic circuit, and the parameter statistic circuit is coupled to the type judging circuit, and the type judging circuit is respectively coupled to the dark portion reinforcing circuit and the hybrid circuit, and the dark portion reinforcing circuit is further coupled The color space conversion module is coupled to the inversion processing module, and the inversion processing module is coupled to the hybrid circuit to reproduce the hybrid circuit and the bright portion. The boosting circuit is coupled. The camera dynamic range expansion device of claim 1, wherein the camera dynamic range expansion device further comprises an automatic gain control circuit coupled to the bright portion boost circuit. The camera dynamic range expansion device according to claim 1, wherein the reverse processing module comprises a 2×2 maximum filling circuit, a 3×3 low-pass filter circuit, and a phase inversion. The processing circuit is configured to couple the color space conversion module to the filling circuit, and the inverting processing circuit is coupled to the hybrid circuit. A camera dynamic range expansion method, the implementation steps include the following: A. Image color space conversion: the image taken by the camera is input into a color space conversion module to convert the color space of the three primary colors of red, green and blue. Color space of brightness and chromatic aberration; B. Image type classification: following the brightness value data of the original image is sent to a parameter statistical circuit for histogram statistical analysis, and then the original image is classified by a type judgment circuit as One of too dark, over bright or extreme; C. Image dark portion reinforcement processing: according to the type judgment circuit, the type of judgment is different, and the image of the over-dark type and the extreme type is sent to a dark portion reinforcing circuit to perform the dark portion reinforcing operation to obtain the dark portion reinforcing image. , then input a hybrid circuit, if it is a super bright image, directly input the hybrid circuit; D. Image inversion processing: the original image data converted by the color space conversion module is simultaneously sent to an inversion Processing the module, performing inversion processing through the inversion processing module to obtain an inverted image to input the hybrid circuit; E. Image mixing processing: continuing to input the inverted image and the dark portion by the hybrid circuit The reinforced image is mixed to obtain a mixed image; F. The image bright portion lifting process is performed by a bright portion lifting circuit to perform a bright portion lifting operation to obtain a high dynamic range image. The camera dynamic range expansion method according to claim 4, wherein the camera dynamic range expansion method further comprises an automatic gain control circuit coupled to the bright portion lifting circuit to increase the brightness of the bright portion. The obtained high dynamic range image is sent to the automatic gain control circuit for contrast adjustment of the brightness and output. The camera dynamic range expansion method according to claim 4, wherein the over-dark, over-bright, and extreme types in the image type classification step are based on two parameter values of Hi and Lo, and the Hi And Lo two parameter value formula is as follows: Where h(i) represents the statistical pixel quantity of the histogram gray scale i, and Th1~Th4 are four custom thresholds, respectively falling in four positions of the histogram gray scale from small to large, when Lo > 2 Hi , the system is too dark, and when Hi > 2 Lo , it is too bright, another or When it is extreme, the P _total is the total number of pixels of the picture. The camera dynamic range expansion method according to claim 6, wherein the type judging circuit further comprises a first multiplexer, a second multiplexer and a third multiplexer coupled to each other, when the type is judged When the circuit judges Lo > 2 Hi , it obtains 2Hi by shifting Hi by 1 bit, and then compares it with Lo. If Lo > 2 Hi is met, the first multiplexer selects 1, and its serial number is [type]. The output 00 represents an over-dark type. If Lo > 2 Hi is not met, the first multiplexer selects 0. At this time, the output is the second multiplexer result, and the second multiplexer determines whether the over-elastic condition is met. Hi > 2 Lo to control, if it is met, the second multiplexer selects 1, the serial number output is 01 for over-bright type, if it does not meet Hi > 2 Lo , the second multiplexer selects 0, at this time, its output For the third multiplexer result, the third multiplexer determines whether it meets the extreme condition or To control, if any of the above two conditions are met, the third multiplexer selects 1. At this time, the serial number output 10 represents the extreme type. If both conditions are not met, the serial number output 11 represents the normal type. The camera dynamic range expansion method according to claim 4, wherein the image mixing processing step performs the multiplication operation on the inverted image and the dark portion image, and the multiplication formula is as follows: y _mix ( i , j )= y _dark ( i , j )× y _inv ( i , j )× k where y _mix is the output value after mixing operation, y _dark is the dark portion to reinforce the image value, and y _inv is the inverse image The image value, k value is the key parameter for outputting a high dynamic range image. The camera dynamic range expansion method according to claim 4, wherein the image type classification step further comprises an operation of a nonlinear curve equation, and the formula of the nonlinear curve equation is as follows: Where y _inv is the inverse value and y _lpf is the input value of the original image data. The camera dynamic range expansion method according to claim 4, wherein the image type classification step further comprises an operation of compensating an equation, and the formula of the compensation equation is as follows: Where y _dark is the dark portion to reinforce the image output value, y _in is the input image grayscale value, and Th _dark is the custom threshold. The camera dynamic range expansion method according to claim 4, wherein the bright portion lifting operation formula of the image bright portion lifting processing step is as follows: Where y _mix is the mixed image value, y _dark is the dark portion to reinforce the image value, Th _b is the threshold value, and y _HDR is the high dynamic range image value. The camera dynamic range expansion method according to claim 11, wherein the dark portion reinforcing circuit further comprises a multiplexer coupled to the temporary storage And the subtractor, when the multiplexer of the dark portion reinforcing circuit receives an over-dark serial number outputted by the type judging circuit, the multiplexer selects 128 as the threshold value output, when the multiplexer of the dark portion reinforcing circuit receives When the type judging circuit outputs an extreme type number, the multiplexer selects 50 as the threshold value output, and the output value is stored in the register, and then the threshold is sent to the subtractor by the register. The image brightness value data input by the type judging circuit is subtracted, and the subtraction result is shifted right by 4 bits and added to the input image brightness value data to obtain a reinforcing operation result. The camera dynamic range expansion method according to claim 11, wherein the bright portion lifting circuit comprises a multiplexer coupled to the multiplexer, a comparator, a subtractor, a multiplier and two adders, The subtractor performs a dark portion enhancement value [dark_enh] input by the dark portion reinforcement circuit minus 50, and the multiplier multiplies the result of the subtractor with itself to obtain a dark portion enhancement square value [dark_enh_mul], and then uses the right shift 8-bit element. Adding a 10-bit right shift to achieve a value multiplied by 0.005, then adding this value to the output value of the mixed image [out_mix], and returning the rear multiplexer [MUX] to select the output.