JPH0844844A - Object detector - Google Patents

Abstract

PURPOSE:To provide an object detector by which a moving area can be accurately detected without any error of detection even at the part of close signal levels. CONSTITUTION:A binary signal to which the differential absolute value between the signal levels of a background image and a present image obtained by a differential means 3 and an absolute value means 4 is processed in a binarizing circuit 11 and the differential absolute value of the gradient between the gradient calculated based on the background obtained by a differential means 9 and an absolute vale means 10 and the gradient calculated based on the present image by calculating the gradient for the respective background image and present image are processed by a logic circuit 13 to detect the object.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an object detecting device for detecting the shape of an object from continuously input moving images.

[0002]

2. Description of the Related Art In recent years, various moving object detecting devices have been developed such as a detecting device for a passing vehicle, a detecting device for a parked vehicle, and a detecting device for an illegal intruder using image processing technology.
Generally, these devices use an input image obtained from an image pickup device such as a video camera as a background image that does not include an object, and take the absolute value of the difference between the images input from the cameras one after another, and change the value first. A region having a change in illuminance is removed from the region having a change by noise removal, labeling, feature extraction, etc., and only the actual object region is extracted.

As a main method for detecting a changed area from a moving image, a method of taking a difference between an input image and an image input a predetermined period (for example, one frame period) before,
The background image can be stored in advance and can be roughly divided into methods for obtaining the difference from the input image.

However, the object detection method based on the difference between the input image and the image input a predetermined time ago cannot detect the object unless the moving object is constantly moving.

On the other hand, the object detection method by taking the difference from the background image can detect even if some object comes into the background image or the object is stationary. Hereinafter, this object detection method will be further described.
Considering the background image signal g1 (i, j) input image g2 (i, j), if the input image is the same as the background image, g1 (i, j) -g2 ( i, j) = 0. On the other hand, when there is an object in the input image that is different from the background image, it is considered that g1 (i, j) -g2 (i, j) ≠ 0 at the pixel (i, j) in which the object is present. Therefore, for all (i, j), object detection is performed with | g1 (i, j) -g2 (i, j) |> k (k is a predetermined threshold value considering noise etc.) be able to.

FIG. 11 is a block diagram of a conventional object detecting device. In the figure, 1 is a signal input of an image (hereinafter referred to as “current image”) that is a processing target of object shape extraction, 2 is a background image signal input, and 3 is a difference between the current image signal 1 and the background image signal input 2. For computing the absolute value circuit for converting the output of the subtractor 3 into an absolute value, and 12 for binarizing the output of the absolute value circuit 4 with a predetermined threshold value. Is.

Next, the operation will be described. FIG. 12 is a waveform diagram for explaining the operation of the conventional object detection device. For example, when FIG. 12A is input to the circuit of FIG. 11 as a background image signal input and FIG. 12B is a luminance signal of one horizontal line of the current image signal input to the circuit of FIG. The difference absolute value of the signal levels of the current image signal and the background image signal shown in c) is obtained. As the absolute difference value of the signal level, when the signal level at the same coordinate between the current image signal and the background image signal changes, a value corresponding to the amount of change is output. By binarizing this difference absolute value output (c) with a predetermined threshold value by the binarizing circuit 12, (d) is obtained. This (d) is detected as a portion where the signal level has changed between the background frame (a) and the current frame (b), that is, a portion where the object exists.

[0008]

Since the conventional object detecting device is configured to judge the object based on the change in the signal level as described above, the part (a) in FIG. As described above, there is a drawback that it is determined that there is no change in a portion where the signal level is close (| g1 (i, j) -g2 (i, j) | <k) even if there is a change in practice. That is, there is a problem in that a detection omission occurs in which a signal is detected when the background and the object are close to each other in signal level, although the object is actually present but the object is not present. More specifically, based on the simulation image of FIG. 13, for example, (A) is a background image and (A) is
Is the current image signal (both are images actually having 256 gradations), the binarized image of the absolute value of the difference between these signal levels is as shown in (c), and the signal level is particularly close to that of the background image. There are many missing detections from the right shoulder to the right wrist.

The present invention has been made in order to solve the above problems, and provides an object detecting device capable of accurately detecting the shape of an object without omission of detection even in a portion where the signal level is close. With the goal.

[0010]

According to a first aspect of the present invention, there is provided an object detecting apparatus comprising: a first absolute difference value calculating means for calculating an absolute difference value between signal levels of a background image and a current image; First binarizing means for binarizing the output of the absolute difference value calculating means with a predetermined threshold, gradient calculating means for calculating a gradient for each of the background image and the current image, and a gradient calculated based on the background image. And a second difference absolute value calculation means for calculating a difference absolute value of the gradient calculated based on the current image, and a second difference absolute value calculation means for binarizing the output of the second difference absolute value calculation means with a predetermined threshold value. Object detection is performed by including a binarization unit and a logic circuit that calculates a logical sum of the output of the first binarization unit and the output of the second binarization unit.

An object detecting apparatus according to a second aspect of the present invention comprises a first difference absolute value calculating means for calculating the difference absolute value of the signal levels of the background image and the current image, and the first difference absolute value calculating means. First binarizing means for binarizing an output with a predetermined threshold value, horizontal gradient calculating means for calculating a horizontal gradient for each of the background image and the current image, and a horizontal gradient calculated based on the background image and the current image. Second difference absolute value calculating means for calculating the difference absolute value of the horizontal gradient calculated based on the above, and second binary value for binarizing the output of the second difference absolute value calculating means with a predetermined threshold value. And a vertical gradient calculating means for calculating a vertical gradient for each of the background image and the current image, and an absolute difference between the vertical gradient calculated based on the background image and the vertical gradient calculated based on the current image. The third difference absolute value calculation means and the third difference Third binarizing means for binarizing the output of the absolute value calculating means with a predetermined threshold, the output of the first binarizing means, the output of the second binarizing means, and the third binary Object detection is performed by using a logic circuit that calculates the logical sum of the outputs of the conversion means.

According to a third aspect of the present invention, there is provided an object detecting apparatus, wherein first difference absolute value calculating means for calculating a difference absolute value between signal levels of a background image and a current image, and a signal for each of the background image and the current image. Gradient calculation means for calculating the gradient of the level; second difference absolute value calculation means for calculating the difference absolute value of the gradient calculated based on the background image and the gradient calculated based on the current image; An addition means for adding the output of the first difference absolute value calculation means and the output of the second difference absolute value calculation means, and a binarization means for binarizing the output of the addition means with a predetermined threshold value are provided. The object is detected.

According to a fourth aspect of the present invention, there is provided an object detecting device, wherein first difference absolute value calculating means for calculating the difference absolute value of the signal levels of the background image and the current image, and the horizontal difference for each of the background image and the current image. Horizontal gradient calculating means for calculating a gradient; vertical gradient calculating means for calculating a vertical gradient for each of the background image and the current image; a horizontal gradient calculated based on the background image and the current image; Second absolute difference value calculating means for calculating the absolute difference value of the horizontal gradient, and the absolute difference value of the vertical gradient calculated based on the background image and the vertical gradient calculated based on the current image. Third difference absolute value calculation means, the first difference absolute value calculated by the first difference absolute value calculation means, the second difference absolute value calculated by the second difference absolute value calculation means, Third absolute difference calculation And adding means for adding a third absolute difference value calculated from the stage, and performs object detection and a binarizing means for binarizing the output of said adding means with a predetermined threshold value.

According to a fifth aspect of the present invention, there is provided an object detecting apparatus, which comprises first difference absolute value calculating means for calculating a difference absolute value between signal levels of a background image and a current image, and a gradient for each of the background image and the current image. Gradient calculating means for calculating
A second difference absolute value calculating means for calculating a difference absolute value of the gradient calculated based on the background image and the gradient calculated based on the current image; and an output of the second difference absolute value calculating means. Multiplying means for multiplying by a predetermined number, adding means for adding the output of the first difference absolute value calculating means and output of the multiplying means, and binarizing means for binarizing the output of the adding means with a predetermined threshold value. And is used to detect an object.

[0015]

In the object detecting apparatus according to the first aspect of the present invention, with the above configuration, the part having a large difference in signal level or the part having a large difference in gradient is determined as the object region.

In the object detecting apparatus according to the second aspect of the present invention, the part having a large difference in signal level and the part having a large difference in gradient in the horizontal direction or the vertical direction are determined as the object region.

In the object detecting apparatus according to the third aspect of the present invention, with the above configuration, not only the signal level value itself but also the difference in gradient is added to determine the object area.

In the object detecting apparatus according to the fourth aspect of the present invention, the object area is determined by adding a difference in both horizontal and vertical gradients to a portion having a large difference in signal level.

In the object detecting apparatus according to the fifth aspect of the present invention, the object area is determined by changing the weighting of the difference in the signal level value itself and the difference in the gradient.

[0020]

【Example】

Example 1. First Embodiment FIG. 1 is a block diagram of an object detection device according to a first embodiment of the present invention. In the figure, 1 is a current image signal input terminal, 2 is a background image signal input terminal, 3 is a current image signal 1
And a background image signal 2 are subtracted, 4 is an absolute value circuit for obtaining the absolute value of the difference value calculated by the subtractor 3, and 5 and 6 are gradients of the signal level of the current image signal 1 (hereinafter, Delay means and subtractor for calculating the "gradient", 7 and 8 delay means and subtractor for calculating the gradient of the background image signal 2, 9 and 10 the gradient of the current image signal and the background image signal Subtractor and absolute value circuit for calculating the difference absolute value of the gradient of, a binarization circuit 11 for binarizing the output of the absolute value circuit 4 with a predetermined threshold value, 12 an output of the absolute value circuit 10. Is a binarization circuit for binarizing the output of the binarization circuit with a predetermined threshold value, and 13 is a logic circuit for calculating the logical sum of the output of the binarization circuit 11 and the output of the binarization circuit 12.

Next, an operation example will be described with reference to the waveform diagrams of FIGS. 2 (a) shows the background image signal input,
It is assumed that (b) is input to the circuit of FIG. 1 as a luminance signal of one horizontal line of the current image signal input. The absolute value of the difference between the signal levels of the current image signal and the background image signal is obtained by the subtractor 3 and the absolute value circuit 4 as in the conventional case. The absolute difference value of this signal level is shown in FIG. In addition to this, for the current image signal input 1, the output of the delay means 5 and the subtractor 6
The gradient is calculated by The gradient of the current image signal, that is, the output of the subtracter 6 is shown in FIG. Similarly, the gradient of the background image signal input 2 is calculated by the output of the delay unit 7 and the subtractor 8. The gradient of this background image signal, that is, the output of the subtracter 8 is shown in FIG.

Next, the subtracter 9 performs difference processing on the gradient (d) of the current image signal output from the subtractor 6 and the gradient (c) of the background image signal output from the subtractor 8. As for the difference value output here, if the current image signal and the background image signal are the same image, the gradient of the current image signal and the gradient of the background image signal are the same value, and the difference result is 0.
Becomes On the contrary, when the change is large, it takes a large positive or negative value. Since only the amount of change is focused here, the absolute value circuit 10 next performs absolute value processing on the gradient difference signal.

The output of the absolute value circuit 10, that is, the difference absolute value of the gradient is shown in FIG. The absolute difference value of the gradient is output when the gradient between the current image signal and the background image signal changes at the same coordinate. The binarization circuit 12 binarizes the absolute difference value (e) with a predetermined threshold value to obtain the binarized output of (f). On the other hand, as in the conventional example, the binarization circuit 11 binarizes the signal level difference absolute value (g) with a predetermined threshold value to obtain a binarized output of (h). Logic circuit 13
Then, when the logical sum of the binarized output of (f) and the binarized output of (h) is taken, the binarized output of (i), that is, the object detection signal is obtained. Here, the threshold values of the binarization circuits 11 and 12 are set to about 10 to 30 with a margin for the noise amount of 5 to 10 caused by sampling, for example, for an image of 256 gradations. .

When (g) of FIG. 2 corresponding to the absolute value of the signal level difference in the conventional example is binarized, as shown in (h), "0" is obtained in the portion (a) where the signal levels are close. Although a part is formed, a binary output without detection omission is obtained as shown in (i) by taking the logical sum of (f) and (h) obtained by binarizing the absolute difference value of the gradient. Further explaining in detail based on the simulation image of FIG. 13, assuming that (A) is a background image and (A) is a current image signal (both actually images having 256 gradations), the difference between these signal levels The logical sum image of the binarized output of the absolute value and the binarized output of the absolute value of the gradient is as shown in (d), and the omission of detection is reduced as compared with (c) of the conventional example.

Example 2. FIG. 3 is a block diagram of an object detection device according to the second embodiment of the present invention. In the figure, 1 is a current image signal input, 2 is a background image signal input, 3 is a subtractor for calculating the difference between the current image signal 1 and the background image signal 2, and 4 is the absolute value of the difference value calculated by the subtractor 3. An absolute value circuit for obtaining a value, 31 and 32 are delay means and subtractors for calculating the horizontal gradient of the current image signal 1, 33 and 34 are delay means and subtractor for calculating the horizontal gradient of the background image signal 2. , 35 and 36 are delay means and subtractors for calculating the vertical gradient of the current image signal 1, 37 and 38 are delay means and subtractors for calculating the vertical gradient of the background image signal 2, and 39 and 40 are current delay elements and subtractors. Subtractors and absolute value circuits for calculating the absolute difference between the horizontal gradient of the image signal and the horizontal gradient of the background image signal, 41 and 42 represent the absolute difference between the vertical gradient of the current image signal and the vertical gradient of the background image signal. Subtractor for calculating And an absolute value circuit, 11 is a binarization circuit for binarizing the output of the absolute value circuit 4 with a predetermined threshold value, and 43 is a binary value for binarizing the output of the absolute value circuit 40 with a predetermined threshold value. Circuit,
Reference numeral 44 denotes a binarization circuit that binarizes the output of the absolute value circuit 42 with a predetermined threshold value, and 13 denotes the output of the binarization circuit 11, the output of the binarization circuit 43, and the output of the binarization circuit 44. It is a logic circuit that calculates a logical sum.

Next, the operation will be described. In FIG. 3, the absolute value of the difference between the signal levels of the current image signal and the background image signal is obtained by the subtracter 3 and the absolute value circuit 4 as in the first embodiment. The difference between the second embodiment and the first embodiment is that
It is to take a logical sum by binarizing the absolute value of the difference between the gradients in both vertical directions. That is, for the current image signal input 1, the output of the delay means 31 and the subtractor 32 calculate the horizontal gradient, and the output of the delay means 35 and the subtractor 36 calculate the vertical gradient. With respect to the background image signal input 2, the output of the delay unit 33 and the subtractor 34 calculate the horizontal gradient, and the output of the delay unit 37 and the subtractor 38 calculate the vertical gradient. Regarding the horizontal gradient of the current image signal output from the subtractor 32 and the horizontal gradient of the background image signal output from the subtractor 34, the subtractor 39 and the absolute value circuit 40
The absolute difference value is calculated by.

Further, for the vertical gradient of the current image signal output from the subtractor 36 and the vertical gradient of the background image signal output from the subtractor 38, a subtractor 41 and an absolute value circuit 42 calculate difference absolute values. To be done. Absolute value circuit 4
The absolute value of the difference between the signal levels calculated in
Then, the difference absolute value of the horizontal gradient calculated by the absolute value circuit 40 is binarized by the binarizing circuit 43, and the difference absolute value of the vertical gradient calculated by the absolute value circuit 44 is binarized by the binarizing circuit 44. An object detection signal is obtained by taking the logical sum of these three binarized outputs in the logic circuit 13.

Next, the effect of calculating the gradient in both the horizontal and vertical directions will be described. It is assumed that an object having a signal level as shown in FIG. 2A is vertically and horizontally translated from the position shown in FIG. 4A to the position shown in FIG. 4B. At this time, if only the absolute difference of the horizontal gradient is added to the absolute value of the difference of the signal level, both the absolute value of the difference of the signal level and the absolute value of the difference of the horizontal gradient are small near the overlapping portion of the object. As shown in 4 (c), detection omission occurs. On the other hand, since the absolute value of the difference between the vertical gradients is large near the overlapping portion of the objects, when the absolute value of the difference between the gradients in both the horizontal and vertical directions is added, FIG.
It is possible to detect an object without omission of detection as in.

Example 3. FIG. 5 is a block diagram of an object detection device according to the third embodiment of the present invention. In FIG.
1 is a current image signal input terminal, 2 is a background image signal input terminal,
3 is a subtracter for calculating the difference between the current image signal 1 and the background image signal 2, 4 is an absolute value circuit for obtaining the absolute value of the difference value calculated by the subtractor 3, and 5 and 6 are gradients of the current image signal 1. , 8 are delay means and subtractors for calculating the gradient of the background image signal 2, and 9 and 10 are absolute differences between the gradient of the current image signal and the gradient of the background image signal. A subtracter and an absolute value circuit for calculating a value, 51 is an adder for adding the output of the absolute value circuit 4 and the output of the absolute value circuit 10, and 52 is a binary value of the output of the adder 51 with a predetermined threshold value. It is a binarization circuit for converting into a binary.

Next, an operation example will be described based on the waveform charts of FIGS. 6 (a) shows the background image signal input,
It is assumed that (b) is input to the circuit of FIG. 5 as a luminance signal of one horizontal line of the current image signal input. The absolute value of the difference between the signal levels of the current image signal and the background image signal is obtained by the subtractor 3 and the absolute value circuit 4 as in the conventional case. The difference absolute value of this signal level is shown in FIG. In addition to this, for the current image signal input 1, the output of the delay means 5 and the subtractor 6
The gradient is calculated by The gradient of this current image signal, that is, the output of the subtracter 6 is shown in FIG. Similarly, the gradient of the background image signal input 2 is calculated by the output of the delay unit 7 and the subtractor 8. The gradient of this background image signal, that is, the output of the subtracter 8 is shown in FIG.

Next, the subtracter 9 performs difference processing on the gradient (d) of the current image signal output from the subtractor 6 and the gradient (c) of the background image signal output from the subtractor 8. As for the difference value output here, if the current image signal and the background image signal are the same image, the gradient of the current image signal and the gradient of the previous image signal have the same value, and the difference result is 0. On the contrary, when the change is large, it takes a large positive or negative value. Since only the amount of change is focused here, the absolute value circuit 10 next performs absolute value processing on the gradient difference signal.

The output of the absolute value circuit 10, that is, the difference absolute value of the gradient is shown in FIG. 6 (e). When the gradient between the current image signal and the background image signal changes at the same coordinate, the difference absolute value of the gradient is output as a value corresponding to the amount of change. The absolute difference (e) of the gradient and the absolute value of the difference (f) of the signal level are added by the adder 51. This added value is as shown in FIG. 6 (h), and an object detection signal is obtained by binarizing the added value with a predetermined threshold value by the binarizing circuit 52. 6 (f) corresponding to a signal level difference absolute value which is a conventional example and FIG. 6 (h) which is the sum of the signal level difference absolute value and the gradient difference absolute value of the third embodiment have the same threshold. When the value is binarized, FIG.
As shown in (g), a part where the signal level is close to (a) has a part that becomes “0”, but in (h), there is no detection omission as shown in FIG. 6 (i). It becomes a digitized output. Further explaining in detail based on the simulation image of FIG. 13, assuming that (A) is a background image and (A) is a current image signal (both actually images having 256 gradations), the difference between these signal levels The binarized image obtained by binarizing the added value of the absolute value and the difference between the absolute value and the gradient is as shown in (e), and the omission of detection is reduced as compared with (c) of the conventional example.

Example 4. FIG. 7 is a block diagram of an object detection device according to the fourth embodiment of the present invention. In FIG.
1 is a current image signal input, 2 is a background image signal input, 3 is a subtractor for calculating the difference between the current image signal 1 and the background image signal 2, and 4 is an absolute value of the difference value calculated by the subtracter 3. Absolute value circuits to be obtained, 31, 32 are delay means and subtractors for calculating the horizontal gradient of the current image signal 1, 33, 34
Is a delay means and subtractor for calculating the horizontal gradient of the background image signal 2, 35 and 36 are delay means and subtractor for calculating the vertical gradient of the current image signal 1, and 37 and 38 are for the background image signal 2. Delay means and subtractors for calculating the vertical gradient, 39 and 40 are subtractors and absolute value circuits for calculating the absolute difference between the horizontal gradient of the current image signal and the horizontal gradient of the background image signal, and 41 and 42 are A subtracter and an absolute value circuit for calculating the difference absolute value between the vertical gradient of the current image signal and the vertical gradient of the background image signal, and 61 indicates the output of the absolute value circuit 4 and the outputs of the absolute value circuit 40 and the absolute value circuit 42. An adder for adding 62 is a binarizing circuit for binarizing the output of the adder 61 with a predetermined threshold value.

Next, the operation will be described. In FIG. 7, the absolute value of the difference between the signal levels of the current image signal and the background image signal is obtained by the subtractor 3 and the absolute value circuit 4 as in the first embodiment. The difference between the fourth embodiment and the third embodiment is that the absolute difference value of the gradient is calculated in both the horizontal and vertical directions and added. That is, for the current image signal input 1, the horizontal gradient is calculated by the output of the delay means 31 and the subtractor 32,
The output of the delay unit 35 and the subtractor 36 calculate the vertical gradient. Also, with respect to the background image signal input 2, the delay means 3
The horizontal gradient is calculated by the output of No. 3 and the subtractor 34, and the vertical gradient is calculated by the output of the delay unit 37 and the subtractor 38.
The absolute difference between the horizontal gradient of the current image signal output from the subtractor 32 and the horizontal gradient of the background image signal output from the subtractor 34 is calculated by the subtractor 39 and the absolute value circuit 40. Further, with respect to the vertical gradient of the current image signal output from the subtractor 36 and the vertical gradient of the background image signal output from the subtractor 38, a subtractor 41 and an absolute value circuit 42 calculate difference absolute values. The absolute difference between the horizontal gradient and the vertical gradient is calculated by the adder 61.
Are both added to the absolute value of the difference between the signal levels.
An object detection signal is obtained by binarizing the output of the adder 61 by a binarizing circuit 62 with a predetermined threshold value.

Next, the effect of calculating the gradient in both the horizontal and vertical directions will be described. Figure 8 (a) for both horizontal and vertical
It is assumed that an object having such a signal level has moved in the vertical direction from the position of FIG. 8A to the position of FIG. 8B. At this time, if only the absolute difference of the horizontal gradient is added to the absolute value of the difference of the signal level, both the absolute value of the difference of the signal level and the absolute value of the difference of the horizontal gradient are small near the overlapping portion of the object. 8 (c) results in omission of detection. On the other hand, since the absolute value of the difference between the vertical gradients becomes large in the vicinity of the overlapped portion of the object, a signal as shown in FIG. 8D is obtained for the gradient components of both the horizontal and vertical terms. When the absolute values of the differences between the horizontal and vertical gradients are added,
An object can be detected without omission of detection as in (e).

Example 5. FIG. 9 is a block diagram of an object detection device according to the fifth embodiment of the present invention. The configuration is such that a multiplier 81 is added after the absolute value circuit 10 of the third embodiment. In FIG. 9, 1 is a current image signal input, 2 is a background image signal input, 3 is a subtracter for calculating the difference between the current image signal 1 and the background image signal 2, and 4 is a difference value calculated by the subtracter 3. Absolute value circuits for obtaining absolute values, 5 and 6 are delay means and subtractors for calculating the gradient of the current image signal 1, and 7 and 8 are delay means and subtractor for calculating the gradient of the background image signal 2. Reference numerals 9 and 10 denote a subtracter and an absolute value circuit for calculating the difference absolute value between the gradient of the current image signal and the gradient of the background image signal. Reference numeral 81 denotes a multiplier for making the difference absolute value of the gradient a predetermined value. Output of absolute value circuit 4 and multiplier 8
An adder for adding the output of 1 and a binarization circuit for binarizing the output of the adder 51 with a predetermined threshold value.

Next, an operation example will be described based on the waveform charts of FIGS. 9 and 10. It is assumed that FIG. 10A is input to the circuit of FIG. 9 as a background image signal input and FIG. 10B is input to the current image signal input as a luminance signal of one horizontal line. The absolute value of the difference between the signal levels of the current image signal and the background image signal is obtained by the subtractor 3 and the absolute value circuit 4 as in the conventional case. The absolute difference value of this signal level is shown in FIG. In addition to this, the gradient of the current image signal input 1 is calculated by the output of the delay means 5 and the subtractor 6. The gradient of this current image signal,
That is, the output of the subtractor 6 is shown in FIG. Similarly, the gradient of the background image signal input 2 is calculated by the output of the delay unit 7 and the subtractor 8. The gradient of this background image signal, that is, the output of the subtracter 8 is shown in FIG.

Next, the subtracter 9 performs difference processing on the gradient (d) of the current image signal output from the subtractor 6 and the gradient (c) of the background image signal output from the subtractor 8. As for the difference value output here, if the current image signal and the background image signal are the same image, the gradient of the current image signal and the gradient of the background image signal are the same value, and the difference result is 0.
Becomes On the contrary, when the change is large, it takes a large positive or negative value. Since only the amount of change is focused here, the absolute value circuit 10 next performs absolute value processing on the gradient difference signal. Absolute value circuit 10
10E, that is, the absolute difference value of the gradient is shown in FIG. When the gradient of the same coordinates between the current image signal and the background image signal changes, a value corresponding to the amount of change is output as the difference absolute value of this gradient.

If the absolute difference (e) of the gradient is amplified by the multiplier 81, for example, twice, the output of the multiplier 81 is shown in FIG.
It becomes like (f). The multiplier output (f) and the signal level difference absolute value (g) are added by the adder 51. This added value is as shown in FIG. 10 (i), and the binarization circuit 12 binarizes it to obtain an object detection signal. 10 (g) corresponding to the signal level difference absolute value which is the conventional example, and (i) which is the sum of the signal level difference absolute value and the gradient difference absolute value which is the present invention is 2 at a predetermined threshold value. When digitized, (g) is
As shown in (h), there is a portion where the signal level is close to (a), which becomes "0". However, in (i),
As shown in (j), the binary output has no detection omission.
At this time, for example, if the multiplier 81 doubles the absolute difference value of the gradient, the added value of the portion (A) becomes larger than that in the third embodiment. Therefore, it is possible to further reduce the omission of detection, or conversely, increase the threshold value to reduce erroneous detection.

[0040]

As described above, according to the first aspect of the present invention, detection omission is eliminated in a portion where the signal level is close, and an object can be detected in conformity with the object shape.

According to the second aspect of the present invention, it is possible to detect an object conforming to the object shape even for various objects of various patterns.

According to the third aspect of the present invention, detection omission is eliminated in a portion where the signal level is close, and the object can be detected in conformity with the object shape.

According to the fourth aspect of the invention, it is possible to detect an object conforming to the shape of the object even with various movements of the object having the pattern.

According to the invention of claim 5, claim 1
According to the invention described above, an object can be detected even in an object with a gentle signal level change that causes detection omission.

[Brief description of drawings]

FIG. 1 is a block diagram of an object detection device according to a first embodiment of the present invention.

FIG. 2 is a waveform diagram for explaining the operation of the object detection device in the first embodiment.

FIG. 3 is a block diagram of an object detection device according to a second embodiment.

FIG. 4 is a waveform diagram for explaining the operation of the object detection device in the second embodiment.

FIG. 5 is a block diagram of an object detection device according to a third embodiment of the present invention.

FIG. 6 is a waveform diagram for explaining the operation of the object detection device in the third embodiment.

FIG. 7 is a block diagram of an object detection device according to a fourth embodiment of the present invention.

FIG. 8 is a waveform diagram for explaining the operation of the object detection device in the fourth embodiment.

FIG. 9 is a block diagram of an object detection device according to a fifth embodiment of the present invention.

FIG. 10 is a waveform diagram for explaining the operation of the object detection device in the fifth embodiment.

FIG. 11 is a diagram for explaining the operation of the conventional object detection device.

FIG. 12 is a block diagram of a conventional object detection device.

FIG. 13 is a simulation image for comparing the conventional example and the object detection device of the example of the present invention.

[Explanation of symbols]

1 current image signal input, 2 previous image signal input, 3 subtractor, 4 absolute value circuit, 5 delay means, 6 subtractor, 7 delay means, 8 subtractor, 9 subtractor, 10 absolute value circuit,
11 binarization circuit, 12 binarization circuit, 51 adder.

Claims (5)

[Claims] 1. A first difference absolute value calculating means for calculating a difference absolute value of a signal level with respect to a background image and an image (hereinafter, referred to as “current image”) which is a processing target of shape extraction of an object, and A first binarization unit that binarizes the output of the first difference absolute value calculation unit with a predetermined threshold value, a gradient calculation unit that calculates a gradient for each of the background image and the current image, and the background image. Second difference absolute value calculating means for calculating the difference absolute value between the gradient calculated based on the current image and the gradient calculated based on the current image, and the output of the second difference absolute value calculating means at a predetermined threshold value. It is provided with a second binarizing means for binarizing, and a logical sum calculating means for calculating a logical sum of the output of the first binarizing means and the output of the second binarizing means. Object detection device. 2. A first difference absolute value calculating means for calculating a signal level difference absolute value between the background image and the current image, and the output of the first difference absolute value calculating means is binary with a predetermined threshold value. First binarizing means for converting the horizontal image and horizontal gradient calculating means for calculating a horizontal gradient for each of the background image and the current image,
A second difference absolute value calculating means for calculating a difference absolute value between the horizontal gradient calculated based on the background image and the horizontal gradient calculated based on the current image; and the second difference absolute value calculating means Second binarizing means for binarizing the output with a predetermined threshold;
Vertical gradient calculating means for calculating a vertical gradient for each of the background image and the current image, and an absolute difference value between the vertical gradient calculated based on the background image and the vertical gradient calculated based on the current image. A third absolute difference value calculating means,
The output of the third absolute difference value calculation means is set to 2 with a predetermined threshold value.
A third binarizing means for binarizing, an OR of calculating the output of the first binarizing means, the output of the second binarizing means and the output of the third binarizing means. An object detection device comprising: an arithmetic unit. 3. A first difference absolute value calculating means for calculating a signal level difference absolute value between the background image and the current image, and a gradient calculating means for calculating a gradient for each of the background image and the current image, Second gradient absolute value calculating means for calculating a gradient absolute value calculated based on the background image and gradient calculated based on the current image, and output of the first differential absolute value calculating means And an adding means for adding the outputs of the second absolute difference value calculating means, and a binarizing means for binarizing the output of the adding means with a predetermined threshold value. 4. A first difference absolute value calculating means for calculating a signal level difference absolute value between the background image and the current image, and a horizontal gradient calculating means for calculating a horizontal gradient for each of the background image and the current image. A vertical gradient calculating means for calculating a vertical gradient for each of the background image and the current image, and a difference absolute value between the horizontal gradient calculated based on the background image and the horizontal gradient calculated based on the current image. Second difference absolute value calculating means for calculating, and third difference absolute value calculating means for calculating the difference absolute value of the vertical gradient calculated based on the background image and the vertical gradient calculated based on the current image A first difference absolute value calculated by the first difference absolute value calculating means, a second difference absolute value calculated by the second difference absolute value calculating means, and a third difference absolute value calculating means The third difference absolute value calculated by And an binarizing unit for binarizing the output of the adding unit with a predetermined threshold value. 5. A first difference absolute value calculation means for calculating a signal level difference absolute value between the background image and the current image, and gradient calculation means for calculating a gradient for each of the background image and the current image, Second gradient absolute value calculating means for calculating a gradient absolute value calculated based on the background image and gradient calculated based on the current image, and output of the second differential absolute value calculating means Is multiplied by a predetermined number, addition means for adding the output of the first absolute difference value calculation means and the output of the multiplication means, and binarization for binarizing the output of the addition means with a predetermined threshold value. An object detection device comprising: