TWI820430B - Device and method for detecting movement of objects in images - Google Patents

Abstract

A device and a method for detecting the movement of an object in an image are provided. The device for detecting the movement of an object in an image includes a weight determination circuit, an image blending circuit, and an object movement detection circuit. The weight determination circuit is used to determine a plurality of weights according to an input image and a background image, each weight corresponding to a pixel position. The image blending circuit is configured to blend the input image and the background image based on the weights to generate an updated background image. The object movement detection circuit is configured to perform a sum of absolute difference (SAD) operation, block by block, on the input image and the background image or updated background image, and generate a moving object indication data, and to generate an object movement signal according to the moving object indication data and at least one threshold. Wherein, each block contains multiple pixels.

Description

Device and method for detecting movement of objects in images

The present invention relates to image processing, and in particular to methods and devices for detecting movement of objects in images.

For security surveillance systems, it is a very important function to accurately detect and determine whether objects in the image are moving. Since security surveillance systems usually find moving objects by comparing the current image with previous images, the way the previous images are generated and the method of judging moving objects greatly affect the performance and sensitivity of the security surveillance system. Conventional security monitoring systems often cause misjudgments (false alarms or missed alarms) because they fail to properly process previous images or the judgment method of moving objects is not accurate enough. Therefore, it is necessary to propose a method and device for detecting the movement of objects in images to improve the reliability of the security monitoring system.

One purpose of the present invention is to provide a device and method for detecting movement of objects in images to improve the deficiencies of the prior art.

An embodiment of the present invention provides a device for detecting movement of objects in an image, including a weighting circuit, an image mixing circuit and an object movement detection circuit. The weight determination circuit is used to determine a plurality of weights based on an input image and a background image, and each weight corresponds to a pixel position. The image mixing circuit is coupled to the weight determination circuit and used to mix the input image and the background image based on the weights to generate an updated background image. The object motion detection circuit is coupled to the image mixing circuit and is used to perform an absolute deviation sum calculation on the input image and the background image or the updated background image block by block, and generate a moving object indication data, and according to the The moving object indication data and at least one threshold generate an object moving signal. Each block contains a plurality of pixels.

Another embodiment of the present invention provides a method for detecting movement of objects in an image, including: determining a plurality of weights based on an input image and a background image, each weight corresponding to a pixel position; and mixing the input image based on the weights. and the background image to generate an updated background image; perform an absolute deviation sum calculation on the input image and the background image or the updated background image block by block, and generate a moving object indication data; and based on the movement The object indication data and at least one threshold generate an object movement signal. Each block contains a plurality of pixels.

The features, implementation and effects of the present invention are described in detail below with reference to the drawings and examples.

The technical terms used in the following description refer to the idioms in the technical field. If some terms are explained or defined in this specification, the explanation or definition of these terms shall prevail.

The disclosure of the present invention includes an apparatus and method for detecting movement of objects in an image. Since some of the components included in the device for detecting movement of objects in images of the present invention may be known components individually, the following description of the known components will not affect the full disclosure and implementability of the device invention. details will be omitted. In addition, part or all of the process of the method for detecting movement of objects in images of the present invention may be in the form of software and/or firmware, and may be executed by the device for detecting movement of objects in images of the present invention or its equivalent device. , without affecting the full disclosure and implementability of the method invention, the following description of the method invention will focus on the step content rather than the hardware.

FIG. 1 is a functional block diagram of an embodiment of a device for detecting movement of objects in an image according to the present invention. The device 100 for detecting movement of objects in an image includes a background image update circuit 110, an object movement detection circuit 120 and a storage circuit 130. The storage circuit 130 may include a register, a random access memory, and/or a buffer. The background image update circuit 110 includes a weight determination circuit 112 and an image mixing circuit 116 . FIG. 2 is a flow chart of an embodiment of a method for detecting movement of objects in an image according to the present invention, which includes the following steps.

Step S210: The weight determination circuit 112 determines a plurality of weights w(x,y) ("(x,y)" represents the coordinates of the pixel) based on the input image IM and the background image BG, and each weight corresponds to a pixel position. The details of step S210 will be explained in detail in FIG. 5 .

Step S220: The image mixing circuit 116 mixes the background image BG and the input image IM based on the weights pixel by pixel to generate an updated background image BG'. The generated updated background image BG' is stored in the storage circuit 130 and will be used as the background image BG for the next round (when there is a new input image IM). More specifically, the image mixing circuit 116 generates the updated background image BG' according to the following equation ( ): (1)

Among them, IM(x,y) is the pixel value of the input image IM at the pixel position (x,y), and BG(x,y) is the pixel value of the background image BG at the pixel position (x,y). In practice, if each pixel value is represented by 8 bits, the values of IM(x,y), BG(x,y), w(x,y) and BG'(x,y) are all between 0~255, and equation (1) can be adjusted to equation (2): (2)

Among them, plus Shifting by another 8 bits represents the rounding operation. A person with ordinary knowledge in the art can understand the implementation details of equation (2) and how to apply equation (1) to the case where the pixel value is not equal to 8 bits, so the details will not be described again.

Step S230: The object motion detection circuit 120 performs a sum of absolute difference (SAD) calculation on the input image and the background image (which can be the background image BG or the updated background image BG') block by block, and generates a Move object instruction data. This step is performed in blocks. Please refer to Figure 3, which shows a schematic diagram of the input image IM and the background image BG (or the updated background image BG'). An image can be a picture or a frame. In the example in Figure 3, an image contains 64 pixels ((x, y), , ) and 16 blocks ((m,n), , ), a block contains 4 pixels. That is to say, the object motion detection circuit 120 performs the SAD operation on each block according to equation (3) (the background image is the background image BG) or equation (4) (the background image is the updated background image BG'). Therefore, A total of 16 SAD values (SAD(m,n), , ). (3) (4)

Among them, bs_x represents the number of pixels of a block on the x-axis, and bs_y represents the number of pixels of a block on the y-axis. That is to say, the size of a block is bs_x*bs_y. In the example of Figure 3, bs_x=bs_y=2.

Then, the object motion detection circuit 120 generates moving object indication data MOI according to all SAD values in the image and the first threshold sad_thd. More specifically, the object motion detection circuit 120 compares SAD(m,n) and the first threshold sad_thd block by block according to equation (5) to obtain the comparison result CR. The comparison results CR of all blocks form a moving object. Indicates the data MOI. (5)

Please refer to Figure 4A. Figure 4A shows an example of the moving object indication data MOI, in which blocks (2,1), (3,1), (2,2), (3,2) and (2,3) The comparison result CR is 1, which means there may be moving objects in this area of the image.

Step S240: The object motion detection circuit 120 generates an object motion signal Sig_alm according to the moving object indication data MOI and at least one threshold. For example, the object motion detection circuit 120 can determine the value of the object motion signal Sig_alm according to Equation (6) or Equation (7). (6) (7)

Among them, N_moi is the number of blocks with a comparison result CR of 1 (in the example of Figure 4A, N_moi=5), and alm_thd1 and alm_thd2 are thresholds. Here, the user can use the size of the set threshold alm_thd1 to indicate that the comparison results are significantly different. This means that a moving object has occurred, and false positives caused by noise can be avoided. Furthermore, the user can use the value of the set threshold alm_thd2 to exclude error messages that may be caused by large changes in brightness caused by switching lights to avoid misjudgments. The object movement signal Sig_alm is 1, which means that the object movement detection circuit 120 has detected a moving object in the input image IM, and the characteristics (such as size) of the moving object meet the user's setting conditions.

Figure 5 is a flowchart of an embodiment of step S210, including steps S510~S590. The weight determination circuit 112 calculates the absolute deviation of the pixel values of the two pixels (one is the target input pixel and the other is the target background pixel) corresponding to a target pixel position (step S580), and calculates the absolute deviation of the pixel value according to the absolute deviation of the pixel value and the weight valve. The value w_thd is used to classify the target pixel position into class A (its weight w (x, y) is the weight w_a of class A) or class B (its weight w (x, y) is the weight w_b of class B) (step S590). The weight determination circuit 112 determines whether to adjust the type A weight w_a (steps S530 to S570) according to whether it operates in the weight adjustment mode (step S520) and whether the input image IM is updated (step S510). Please refer to Figure 3. The target pixel position refers to a set of (x, y) values (for example, the pixel position (4,1) marked in black in Figure 3), and the target input pixel is the pixel of the input image IM at that position (its The pixel value is IM(x,y), the example shown in Figure 3 is IM(4,1)), and the target background pixel is the pixel of the background image BG at that position (its pixel value is BG(x,y), The example shown in Figure 3 is BG(4,1)).

The details of steps S510~S590 are as follows.

Step S510: The weight determination circuit 112 determines whether the input image IM is a new input image. Specifically, the weight determination circuit 112 determines whether the currently input pixel data belongs to a new input image, and then determines whether the currently input pixel data corresponds to the first pixel of the input image IM. In some embodiments, the weight determination circuit 112 determines whether the end of the previous input image IM and/or the beginning of the new input image IM has been reached by counting the number of pixels. If the result of step S510 is yes (ie, the input image has been updated), the weight determination circuit 112 executes step S520; otherwise, the weight determination circuit 112 executes step S580.

Step S520: The weight determination circuit 112 determines whether the background image update circuit 110 or itself operates in the weight adjustment mode. If the result of step S520 is yes (ie, in the weight adjustment mode), the weight determination circuit 112 adjusts the class A weight w_a according to the default value w_a_dft and whether the reset signal w_rst is received (steps S530 to S570); otherwise, the weight The decision circuit 112 executes step S570.

Step S530: The weight determination circuit 112 determines whether the reset signal w_rst is received. If the result of step S530 is yes, the weight determination circuit 112 executes step S540; otherwise, the weight determination circuit 112 executes step S550. When and how to generate the reset signal w_rst will be detailed below in conjunction with Figure 10.

Step S540: The weight determination circuit 112 resets the type A weight w_a to the initial value w_a_ini, and then executes step S580.

Step S550: The weight determination circuit 112 adjusts the Class A weight. For example, the weight determination circuit 112 subtracts a step value (step value) from the Class A weight w_a each time step S550 is executed.

Step S560: The weight determination circuit 112 determines whether the adjusted Class A weight w_a is less than the preset value w_a_dft. The default value w_a_dft can be stored in the storage circuit 130 . If the result of step S560 is yes, the weight determination circuit 112 executes step S570; otherwise, the weight determination circuit 112 executes step S580.

Step S570: The weight determination circuit 112 uses the preset value w_a_dft as the type A weight w_a.

Step S580: The weight determination circuit 112 calculates the absolute deviation Px_d of pixel values between the target input pixel and the target background pixel ( ).

Step S590: The weight determination circuit 112 selects the A-type weight w_a or the B-type weight w_b as the weight w(x, y) corresponding to the target pixel position according to the pixel value absolute deviation Px_d and the weight threshold w_thd. In some embodiments, when the absolute deviation Px_d of the pixel value is greater than the weight threshold w_thd, the target pixel position is classified into category A, which means that the change in the target pixel position between the input image IM and the background image BG is relatively large (possibly Caused by moving objects); when the absolute deviation of the pixel value Px_d is not greater than the weight threshold w_thd, the target pixel position is classified as category B, which means that the change in the target pixel position between the input image IM and the background image BG is relatively small.

In some embodiments, the preset value w_a_dft is smaller than the class B weight w_b. For example, the default value w_a_dft can be 0.25 (or 64, when the pixel value is expressed in 8 bits), and the type B weight w_b can be 0.5 (or 128, when the pixel value is expressed in 8 bits) .

In some cases, when the first input image IM (which will become the first background image BG) after the device 100 for detecting movement of objects in an image is started contains moving objects, the first input image IM in the background The weight should be set lower (that is, the weight of the new input image IM should be set higher) to avoid moving objects remaining in the background (which may cause misjudgment by the device 100 that detects the movement of objects in the image). The purpose of the weight adjustment mode (ie, steps S530~S570) is to handle the above situation. The initial value w_a_ini in step S540 is greater than the preset value w_a_dft in steps S560 and S570. For example, the initial value w_a_ini can be 128, and the default value w_a_dft can be 32; in this case, after the device 100 for detecting the movement of objects in the image is started or the weight determination circuit 112 receives the reset signal w_rst, the weight The weight assigned by the decision circuit 112 to the Class A pixel position (ie, the Class A weight w_a) will initially be equal to 128 (ie, the initial value w_a_ini), and then gradually become smaller until it is equal to 32 (ie, the default value w_a_dft). In another embodiment, when the weight adjustment mode is turned on, for the first input image IM after the device 100 is started or the first input image IM after receiving the reset signal w_rst, the type A weight w_a can be set is the initial value w_a_ini. After that, every time an input image IM is updated (that is, a new input image IM), a fixed value (for example, 16) is deducted from the current Class A weight w_a as the adjusted Class A weight w_a. After a specific number of adjustments (such as 6 times), the weight w_a of type A remains fixed. In the previous embodiment, the operation of the weight adjustment mode is explained by using the weight w_a of type A to gradually become smaller. In implementation, the weight w_b of type B can also be gradually increased to achieve a similar effect. Generally speaking, after the device 100 for detecting the movement of objects in the image is activated or the weight determination circuit 112 receives the reset signal w_rst, when the weight adjustment mode is turned on, there will be a period of time for the weight determination circuit 112 to adjust the Class A weight. w_a or class B weight w_b.

FIG. 6 is a functional block diagram of an embodiment of the object motion detection circuit 120. The object motion detection circuit 120 includes a calculation circuit 610, a comparison circuit 620 and a judgment circuit 630. FIG. 7 shows sub-steps S710 and S720 of step S230, and shows sub-step S730 of step S240.

Step S710: The calculation circuit 610 calculates the sum of absolute deviations SAD(m,n) between the input image IM and the background image block by block. Please refer to equation (3) or (4) and the description of step S230.

Step S720: The comparison circuit 620 generates the moving object indication data MOI according to the absolute deviation sum and the first threshold sad_thd. Please refer to Equation (5), Figure 4A, and related descriptions.

Step S730: The judgment circuit 630 processes the moving object indication data to generate a statistical value, and generates an object moving signal Sig_alm based on the comparison result between the statistical value and the second threshold alm_thd1 and/or the third threshold alm_thd2. More specifically, the judgment circuit 630 counts the number of blocks in the moving object indication data MOI for which the comparison result CR is 1. Taking FIG. 4A as an example, the judgment circuit 630 obtains a statistical value of 5. Next, the judgment circuit 630 compares the statistical value with the second threshold alm_thd1, or compares the statistical value with the second threshold alm_thd1 and the third threshold alm_thd2. For details, please refer to equation (6) or (7) and steps Instructions for S240.

FIG. 8 is a functional block diagram of another embodiment of the object motion detection circuit 120. The object motion detection circuit 120 includes a calculation circuit 610, a comparison circuit 620, a judgment circuit 630, an operation circuit 810, an area selection circuit 820 and a light switching detection circuit 830. Figure 9 shows the sub-steps S910~S930 of step S240.

Step S910: The computing circuit 810 performs a morphology opening operation on the moving object indication data MOI. In addition to making the edges of the moving object indication data MOI smoother, the morphological turning on operation can also remove smaller moving objects in the moving object indication data MOI and/or make the boundaries between moving objects clearer. for clarity.

Step S920: The region selection circuit 820 selects a part of the moving object indication data MOI based on a Region Of Interest (ROI). More specifically, the region selection circuit 820 uses a mask to intersect with the moving object indication data MOI to obtain a portion of the moving object indication data MOI. For example, please refer to FIG. 4B , which shows a schematic diagram of a mask according to an embodiment. As shown in Figure 4B, the region of interest is represented by 1 (i.e., the values of block(1,1), block(2,1), block(1,2), and block(2,2) are 1, the value of other blocks is 0); in other words, the area of interest is the middle part of the image. As shown in FIGS. 4A and 4B , the area selection circuit 820 processes the moving object indication data MOI in units of blocks.

Step S930: The judgment circuit 630 processes the part of the moving object indication data to generate a statistical value, and generates the object moving signal Sig_alm based on the comparison result between the statistical value and the second threshold alm_thd1 and/or the third threshold alm_thd2. This step is similar to step S730, except that the judgment circuit 630 only processes the part of the moving object indication data MOI corresponding to the area of interest in step S930.

The light switching detection circuit 830 will be explained with reference to FIG. 10 . Figure 10 is a flow chart of one embodiment of light switching for image detection according to the present invention, including the following steps.

Step S1010: The light switching detection circuit 830 calculates the current light switching indication value LSD (LSD(q)) based on the sum of absolute deviations (ie, SAD(m,n)), as shown in equation (8). (8)

Among them, bn_x is the number of blocks on the x-axis, and bn_y is the number of blocks on the y-axis. For the example of Figure 4A, bn_x=bn_y=4. In some embodiments, k=1 represents block (0,0), k=2 represents block (1,0), k=3 represents block (2,0), and k=4 represents block (3 ,0), k=5 represents block (0,1), and so on.

Step S1020: After generating LSD(q), the light switching detection circuit 830 stores LSD(q) into the storage circuit 130.

Step S1030: The light switching detection circuit 830 compares the current light switching indication value (LSD(q)) with the previous light switching indication value (LSD(q-1), stored in the storage circuit 130) to generate a light switching indication. Signal L_ind. More specifically, the light switching detection circuit 830 determines the light switching indication signal L_ind according to equation (9), where lsd_thd is the threshold. (9)

It can be seen from equation (8) that the light switching indication value LSD represents the global SAD of an image (ie, the entire image). When the light in the image suddenly changes significantly (for example, the lights in the monitored area are turned on/off), there will be a large difference between LSD(q) and LSD(q-1). In other words, one of the uses of the light switching indication signal L_ind is to indicate the on/off of the light.

Referring to FIG. 8 , the judgment circuit 630 may further determine whether to issue the object movement signal Sig_alm based on the light switching indication signal L_ind. In some embodiments, when the light switching indication signal L_ind is 1, the judgment circuit 630 does not send the object movement signal Sig_alm or sets the object movement signal Sig_alm equal to 0 to avoid mistaking the light switching as a moving object.

In some embodiments, the light switching indication signal L_ind can be used as or used to generate the aforementioned reset signal w_rst. That is, when the light is switched, the weight determination circuit 112 resets the type A weight w_a (step S540).

In some embodiments, the light switching indication signal L_ind can also be used to reset the background image BG. Specifically, when the light is switched, the device 100 deletes the background image BG in the storage circuit 130 (ie, deletes the historical background image) based on the light switching instruction signal L_ind generated by the light switching detection circuit 830, and can reset the system. The count value corresponding to the input image IM causes the device 100 to use the current input image IM as the first background image BG (that is, similar to restarting). Deleting historical background images can prevent the device 100 from causing misjudgments due to light switching.

In some embodiments, the aforementioned equation (8) can be modified into equation (10) to reduce the amount of data (reduce the usage of the storage circuit 130), where . (10)

Those with ordinary skill in the art can design the judgment circuit 630, the operation circuit 810 and the light switching detection circuit 830 based on the above disclosure. That is to say, the judgment circuit 630, the operation circuit 810 and the light switching detection circuit 830 can be Application Specific Integrated Circuit (ASIC) may be implemented by circuits or hardware such as Programmable Logic Device (PLD).

The device and method for detecting movement of objects in an image of the present invention perform pixel-by-pixel operations on the image to update the background image, and perform block-by-block operations on the image to determine whether there are moving objects. Compared with traditional technology, the present invention is less likely to cause misjudgment. In addition, adjusting the weights over time can make the security monitoring system implementing the present invention have higher reliability.

In some embodiments, the order of the steps mentioned in the above-mentioned flow chart can be adjusted according to the actual operation, and can even be executed simultaneously or partially simultaneously.

Although the embodiments of the present invention are described above, these embodiments are not intended to limit the present invention. Those skilled in the art may make changes to the technical features of the present invention based on the explicit or implicit contents of the present invention. All these changes may fall within the scope of patent protection sought by the present invention. In other words, the patent protection scope of the present invention must be determined by the patent application scope of this specification.

100: Device for detecting movement of objects in images 110: Background image update circuit 112: Weight decision circuit 116:Image hybrid circuit 120: Object motion detection circuit 130:Storage circuit IM: input image BG: background image BG':Updated background image SAD(m,n):SAD value sad_thd: first threshold MOI: moving object indication data CR: comparison result Sig_alm: object movement signal w_thd: weight threshold w_a: Class A weight w_a_dft:Default value w_a_ini: initial value w_b: Class B weight w_rst: reset signal 610: Calculation circuit 620: Comparison circuit 630: Judgment circuit alm_thd1: second threshold alm_thd2: third threshold 810: Arithmetic circuit 820: Area selection circuit 830: Light switching detection circuit LSD(q), LSD(q-1): light switching indication value lsd_thd:threshold L_ind: light switching indicator signal step

Figure 1 is a functional block diagram of an embodiment of a device for detecting movement of objects in an image according to the present invention; Figure 2 is a flow chart of one embodiment of a method for detecting movement of objects in an image according to the present invention; Figure 3 shows a schematic diagram of the input image IM and the background image BG (or the updated background image BG'); Figure 4A shows an example of moving object indication data MOI; Figure 4B shows a schematic diagram of an embodiment of the mask; Figure 5 is a flow chart of an embodiment of step S210 in Figure 2; Figure 6 is a functional block diagram of an embodiment of the object motion detection circuit 120; Figure 7 shows the sub-steps of step S230 and the sub-steps of step S240; Figure 8 is a functional block diagram of another embodiment of the object motion detection circuit 120; Figure 9 shows the sub-steps of step S240; and FIG. 10 is a flow chart of one embodiment of light switching for image detection according to the present invention.

100: Device for detecting movement of objects in images

110: Background image update circuit

112: Weight determination circuit

116:Image hybrid circuit

120: Object motion detection circuit

130:Storage circuit

BG: background image

BG':Updated background image

IM: input image

Sig_alm: object movement signal

w_a_dft:Default value

w_a_ini: initial value

w_b: Class B weight

w_rst: reset signal

w_thd: weight threshold

Claims (18)

A device for detecting movement of objects in an image, including: a weight determination circuit for determining a plurality of weights based on an input image and a background image, each weight corresponding to a pixel position; an image mixing circuit for determining based on the Mixing the input image and the background image with these weights to generate an updated background image; and an object motion detection circuit for performing an absolute block-by-block comparison between the input image and the background image or the updated background image. Deviation summation operation, and generating a moving object indication data, and generating an object movement signal according to the moving object indication data and at least one threshold; wherein each block includes a plurality of pixels; wherein the input image includes a target located at A target input pixel at a pixel position. The background image includes a target background pixel at the target pixel position. The weight determination circuit determines based on a weight threshold and an absolute deviation of pixel values between the target input pixel and the target background pixel. Using a first weight as the weight corresponding to the target pixel position, the weight determination circuit further adjusts the first weight when the input image is updated. The device of claim 1, wherein the object motion detection circuit includes: a calculation circuit for calculating the sum of a plurality of absolute deviations between the input image and the background image or the updated background image block by block; a A comparison circuit used to generate the moving object indication data based on the sum of the absolute deviations and a first threshold; and A judgment circuit is used to process the moving object indication data to generate a statistical value, and generate the object moving signal based on a comparison result between the statistical value and a second threshold. The device of claim 2, wherein the judgment circuit generates the object movement signal based on a comparison result of the statistical value being greater than the second threshold and less than a third threshold. The device of claim 1, wherein the object motion detection circuit includes: a calculation circuit for calculating the sum of a plurality of absolute deviations between the input image and the background image or the updated background image block by block; a A comparison circuit used to generate the moving object indication data based on the sum of the absolute deviations and a first threshold; an arithmetic circuit used to perform a morphological opening operation on the moving object indication data; and a judgment circuit used to Process the moving object indication data after the morphology is turned on to generate a statistical value, and generate the object moving signal based on a comparison result between the statistical value and a second threshold. The device of claim 1, wherein the object motion detection circuit includes: a calculation circuit for calculating the sum of a plurality of absolute deviations between the input image and the background image or the updated background image block by block; a a comparison circuit for generating the moving object indication data based on the sum of the absolute deviations and a first threshold; a region selection circuit for selecting a part of the moving object indication data based on a region of interest; and A judgment circuit is used to process the selected part of the moving object indication data to generate a statistical value, and generate the object moving signal based on a comparison result between the statistical value and a second threshold. The device of claim 1, wherein the weight determination circuit has a period of time to adjust the first weight after receiving a reset signal. The device of claim 6, wherein when the adjusted first weight is less than a preset value, the weight determination circuit uses the preset value as the first weight. The device of claim 1, wherein the object motion detection circuit generates a light switching instruction signal based on the absolute deviation sum operation, and the weight determination circuit resets the first weight according to the light switching instruction signal. The device of claim 1, wherein the object motion detection circuit generates a plurality of absolute deviation sums when performing the absolute deviation sum calculation, and the object motion detection circuit calculates a current light switching indication value based on the absolute deviation sums, the The object motion detection circuit further generates a light switching indication signal based on the current light switching indication value and a previous light switching indication value. A method for detecting movement of objects in an image, including: (A) determining a plurality of weights based on an input image and a background image, each weight corresponding to a pixel position; (B) mixing the input image and the input image based on the weights background image to generate an updated background image; (C) perform an absolute deviation sum calculation on the input image and the background image or the updated background image block by block, and generate a moving object indication data; and (D) based on the moving object indication data and at least one valve The value generates an object movement signal; wherein each block includes a plurality of pixels; wherein the input image includes a target input pixel located at a target pixel position, and the background image includes a target background pixel located at the target pixel position, Step (A) includes: (A1) determining a first weight as the weight corresponding to the target pixel position based on a weight threshold and an absolute deviation of pixel values between the target input pixel and the target background pixel; and ( A2) Adjust the first weight when the input image is updated. Such as the method of claim 10, wherein step (C) includes: calculating a plurality of absolute deviation sums between the input image and the background image or the updated background image block by block; and based on the absolute deviation sums and a The first threshold generates the moving object indication data. The method of claim 11, wherein step (D) includes: (D1) processing the moving object indication data to generate a statistical value; and (D2) generating the object based on a comparison result between the statistical value and a second threshold mobile signal. The method of claim 12, wherein step (D2) includes generating the object movement signal based on a comparison result of the statistical value being greater than the second threshold and less than a third threshold. The method of claim 11, wherein step (D) includes: selecting a portion of the moving object indication data based on a region of interest; Processing the selected part of the moving object indication data to generate a statistical value; and generating the object moving signal based on a comparison result between the statistical value and a second threshold. The method of claim 10, wherein the step (A2) includes adjusting the first weight within a period of time after receiving a reset signal. The method of claim 10 further includes: generating a light switching indication signal based on the absolute deviation sum operation; and resetting the first weight according to the light switching indication signal. Such as the method of claim 10, wherein step (C) is to generate a plurality of absolute deviation sums, and the method further includes: calculating a current light switching indication value based on the absolute deviation sums; and based on the current light switching indication value and A previous light switching indication value generates a light switching indication signal. The method of claim 10 further includes: resetting the background image according to a light switching instruction signal.

Images