JP2010147546A - Image processor - Google Patents

Abstract

In conventional image processing, if the amount of movement is large in adjusting the focal position of a camera from an object to be imaged, the shift from the optimum position is large, and if it is small, it takes time to specify the position. In addition, since the focus matching evaluation is based on a large pattern contour waveform, it does not match a small pattern.
A first detection target is detected by a first detection unit and a second detection target smaller than the first detection target is detected by a second detection unit from an imaged input image of the imaging unit. The first and second detection objects are recognized by the recognition means, and the degree of coincidence evaluation of the focus of the imaging means with respect to the first and second detection objects from the first and second detection objects and the input image The evaluation value is presented by the focus evaluation means, the range of the focus setting value is determined from the evaluation value of the first detection object by the first focus adjustment means, and within the focus setting value range by the second focus adjustment means, The focus setting value is determined from the evaluation value of the second detection object, and the focus setting value is changed by a specified amount by the focus changing means.
[Selection] Figure 1

Description

  The present invention relates to an image processing apparatus that processes an image in order to recognize an object from an image captured by an imaging unit such as a camera, and in particular, adjusts the setting values of the focus and zoom position of the imaging unit such as a camera to an optimal state. About things.

  In an image recognition apparatus, for example, an apparatus disclosed in Japanese Patent Application Laid-Open No. 2006-235921 adjusts a camera setting value (focus or zoom position) to an optimal state from an image in which an object is captured. In this technique, the focus position of the electric lens of the camera is sequentially moved to a predetermined position, and the license plates of a plurality of vehicles are detected for each moved focus position to express the degree of blur of the serial number of the license plate. An evaluation value is obtained, a representative evaluation value for each focus position is calculated using the obtained plate position and evaluation value, and a focus position that is in best focus is specified. Here, the evaluation value indicating the degree of blur is a contour waveform width indicating the sharpness of the edge. As other evaluation values, a serial number, a transportation branch code (a land transportation branch code), a certainty of a vehicle type code or a use code (a value representing recognition accuracy) may be used.

JP 2006-235922 A

  The conventional image processing apparatus has a problem that it is difficult to determine the movement amount when moving the focus position. If this amount of movement is increased, the deviation from the optimum focus position may be increased. On the other hand, if the amount of movement is reduced, it will take a long time to specify the focus position. In addition, since the evaluation value for determining whether or not the focus is in focus is obtained from the contour waveform of the largest pattern in the license plate, which is the serial number, the smaller the character, the greater the influence of the focus shift. For this reason, there is a problem that this evaluation value is smaller than the serial number and does not necessarily match a complicated transportation branch code. Furthermore, when the light is reflected on the character pattern, a strong edge is generated due to the reflected light, and there is a problem that the outline waveform and the certainty cannot properly represent the focus shift.

  The present invention has been made to solve the above-described problem, and even when there are a plurality of different objects, it is possible to obtain a setting value of an imaging unit such as an optimal camera at high speed, It is another object of the present invention to obtain an image processing apparatus that can obtain an imaging means setting value such as an appropriate camera even when there is reflected light on an object.

  The image processing apparatus according to the present invention includes an imaging unit that captures an image and uses it as an input image, a first detection unit that detects a first detection object from the input image, and a first detected by the first detection unit. Second detection means for detecting a second detection object smaller than the detection object, a first detection means, a recognition means for recognizing the detection object detected by the second detection means, Focus evaluation means for obtaining a focus evaluation value indicating a degree of focus coincidence with the detection target of the imaging means from the detection results and input images of the detection means and the second detection means, and a plurality of images obtained by sequentially changing the focus setting value greatly A first focus adjustment unit that determines a range of the focus setting value from the focus evaluation value of the first detection target in the input image, and a focus setting within the range determined by the first focus adjustment unit A second focus adjustment unit that determines a focus setting value from a focus evaluation value of a second detection target in a plurality of input images that are picked up in order, and a change in the amount in which the focus setting value of the imaging unit is determined Focus changing means.

  According to the image processing apparatus of the present invention, the first detection object detected by the first detection means and the first detection object detected by the second detection means from the input image taken by the imaging means. The second detection object smaller than the first detection object is recognized by the recognition means, and the focus of the degree of coincidence evaluation of the imaging means with respect to the first and second detection objects from the first and second detection objects and the input image The evaluation value is presented by the focus evaluation means, the first focus adjustment means determines the range of the focus setting value from the evaluation value of the first detection object, the second focus adjustment means within the focus setting value range, Since the focus setting value is determined from the evaluation value of the detection object 2 and the focus setting value is changed by the designated amount by the focus changing means, the optimum focus setting for the second detection object that is a small pattern such as a small character is performed. It is possible to determine the value in a short time. In addition, by performing adjustment processing while switching the target for which the focus evaluation value is to be obtained, even when there are multiple types of detection / recognition targets, it is possible to specify a focus setting value that is in focus overall. Have.

Embodiment 1 FIG.
Embodiment 1 of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic configuration diagram showing Embodiment 1 of the present invention. In the figure, 1 is an image pickup means for picking up an image and using it as an input image, 2 is a first detection means for detecting a first detection object from the input image, and 3 is a position of the object detected by the first detection means. Second detection means for detecting a second detection object having a shape smaller than that of the first detection object from the input image based on the input image, and 4 indicates detection by the first detection means 2 and the second detection means 3. The recognizing means 5 for recognizing the detected object 5 indicates how well the imaging means 1 is focused on the detected object from the detection results of the first detecting means 2 and the second detecting means 3 and the input image. A focus evaluation means 7 for obtaining a focus evaluation value to be shown is a first 7 for determining a range of the focus setting value from the focus evaluation values of the first detection object in a plurality of input images picked up by sequentially changing the focus setting value. Focus adjustment means Reference numeral 8 denotes a first focus setting value that is determined from the focus evaluation values of the second detection target in a plurality of input images that are picked up by sequentially changing the focus setting value within the range determined by the first focus adjusting means 7. A focus adjustment unit 2 and a focus change unit 6 change the focus setting value of the imaging unit 1 by a specified amount.

  Next, the operation will be described with reference to the drawings. FIG. 2 is a flowchart of the adjustment process in the first embodiment of the present invention. Here, the operation will be described with reference to FIGS. 3 is a flowchart showing the processing procedure of the first stage of focus adjustment, FIG. 4 is a flowchart showing the processing procedure of the second stage of focus adjustment, FIG. 5 is a diagram showing display characters on the license plate, and 9 is a number. Plate 10 is a serial number, 11 is a transportation branch code, 12 is a vehicle type code, and 13 is a use code. FIG. 6 is a diagram showing the detection results of the first detection means 2, and 14 to 17 are detection positions. FIG. 7 is a diagram showing the result of the first stage of focus adjustment, where 18 is the minimum value of the focus setting value range and 19 is the maximum value of the focus setting value range. FIG. 8 is a diagram showing a detection result of the second detection means 3, and 20 to 24 are detection positions.

  In this embodiment, an example is shown in which a license plate is a processing target and adjustment processing is performed using characters on the license plate. The first detection object will be described as a serial number 10, and the second detection object will be described as a transportation branch code 11, a vehicle type code 12, and a use code 13. In the focus adjustment, the first focus adjustment means 7 performs the focus adjustment first stage processing (step S1), and then the second focus adjustment means 8 performs the focus adjustment second stage processing (step S2) 2. Consists of stages. First, the details of the focus adjustment first stage process will be described with reference to the flowchart of FIG.

  First, the first focus adjustment unit 7 initializes the focus value to a predetermined value (step SA1). For example, the smallest value that can be set. Thereafter, the first focus adjustment means 7 resets the counter and timer (step SA2). The counter counts the number of times the first detection object is detected, and the timer is for knowing the elapsed time after the reset point.

  Next, the image pickup means 1 picks up an image to be input (step SA3). This input image is, for example, 8-bit digital data for each pixel (a monochrome image in which each pixel takes a value of 0 to 255), and is stored in a data holding means (not shown) such as a memory. Thereafter, the first detection means 2 performs a first detection process to detect each pattern of the serial number, and the recognition means 4 performs recognition (step SA4). As a serial number detection / recognition method, the method described in “Mitsubishi Electric Technical Report Vol. 62 No. 2 pp. 9-12“ Number Plate Recognition Technology ”” is used.

  When the detection is successful in step SA4 (step SA5), the focus evaluation means 5 calculates a focus evaluation value for the image at the detection position of the serial number (step SA6). Since there are four serial numbers, detection positions 14 to 17 are obtained as detection results as shown in FIG. 6, and focus evaluation values corresponding to these positions are calculated. For example, the focus evaluation value is calculated by subtracting the minimum pixel value from the maximum pixel value in the region. Alternatively, edge extraction may be performed and an average value of edge strengths may be used as the evaluation value. These values tend to decrease as the image is blurred. Also, a contour waveform width indicating edge sharpness may be used as in the conventional case. In this case, when the image is blurred, it tends to increase. For convenience of description of the subsequent processing, in this embodiment, the focus evaluation unit 5 outputs a focus evaluation value that tends to decrease as the image is blurred. For example, when using an index that tends to become a large value when the image is blurred, such as the contour waveform width described above, conversion is performed such that the obtained contour waveform width or the like is subtracted from a constant to obtain a focus evaluation value. And

  The first focus adjusting means 7 next checks the elapsed time (step SA7) and then checks the number of detections (step SA8). If the focus setting value deviates greatly from the desired range, it cannot be detected at all. However, this is determined in step SA7, and when a predetermined time has elapsed, the loop of steps SA3 to SA8 is exited to correspond to the focus setting value. The comprehensive evaluation value is calculated by the first focus adjustment means 7 (step SA9). Similarly, when the detection is performed a predetermined number of times, the process exits the loop at step SA8 and proceeds to step SA9.

As a comprehensive evaluation value calculation method in step SA9, for example, if the number of detections has not reached a predetermined number, a constant (0 in this embodiment) is used. The average value of the focus evaluation values obtained is taken as the overall evaluation value. Alternatively, the target focus evaluation value may be narrowed down based on the detection / recognition result, and calculation may be performed within the narrowed range. As a narrowing-down method, for example, there are the following methods (1) to (3). In general, it is difficult to compare the degree of focus between images with different conditions (for example, those with a license plate in the distance and those in the vicinity, those with different objects, etc.) More stable adjustment can be performed by narrowing down images under the same conditions and obtaining a comprehensive evaluation value.
(1) Narrow down the coordinates of the detection position to those within the predetermined range (2) Narrow down the detected object to the size within the predetermined range (3) Narrow down the recognition result to the predetermined range ( (For example, if it is a transportation branch code, limit it to “Shonan”)

  Thereafter, the focus changing means 6 changes the focus setting value by a predetermined width (SA10). For example, when the smallest value that can be set in step SA1 is set, a positive constant (hereinafter, this constant is referred to as a first focus setting movement width) so that the setting value gradually increases. ) Is added. Since the serial number is a relatively large character pattern, even if the first focus setting movement width is set to a large value, the focus adjustment first stage can be achieved in a short time by setting a large movement width without missing the preferred focus setting value range. Can finish. If the focus setting value after the change is within the valid range, for example, if it is less than the maximum settable value, the process returns to step SA2 and repeats the process (step SA11). The adjusting means 7 determines the focus setting value range (step SA12) and ends the focus adjustment first stage process. As a method for determining the focus setting value range, for example, a minimum value and a maximum value of a focus setting value at which the overall evaluation value is equal to or greater than a threshold value are obtained and set as the focus setting value range. FIG. 7 is a graph showing this example. When the value obtained in step SA9 is plotted with the overall evaluation value on the vertical axis and the focus setting value on the horizontal axis, the focus corresponding to the overall evaluation value that is equal to or greater than the threshold value. A minimum value 18 and a maximum value 19 of the set value are obtained, and this is the set value range.

  Next, details of the focus adjustment second stage process will be described with reference to the flowchart of FIG. First, the second focus adjustment means 8 initializes the focus setting value according to the focus setting value range obtained in the first stage of focus adjustment (step SB1). For example, the minimum value of the focus setting value range. Thereafter, the second focus adjustment unit 8 resets the counter and the timer (step SB2) in the same manner as steps SA2 to SA4 in the first stage of focus adjustment, and after the image pickup unit 1 performs the image pickup (step SB3). The first detection means 2 performs the first detection process to detect each pattern of the serial number, and the recognition means 4 recognizes it (step SB4).

  Thereafter, the second detection means 3 performs a second detection process, and based on the result of the first detection process, each pattern of the transportation branch code, the vehicle type code, and the application code is detected, and the recognition means 4 Is recognized (step SB5). As the second detection / recognition processing method, for example, the method described in the above-mentioned “Mitsubishi Electric Technical Report Vol.62 No.2 pp.9-12“ Number plate recognition technology ”” is used. When the detection is successful (step SB6), the focus evaluation means 5 calculates a focus evaluation value for the image at the detection position of the second detection target (step SB7), and regions 20 to 24 in FIG. The target of step SB7 may be all of the second detection target, or may be limited to the transportation branch code, etc. Further, calculation of the focus evaluation value The method may be the same method as step SA6 in the first stage of focus adjustment, or another method may be used.

  After that, the second focus adjustment means 8 checks the elapsed time (step SB8) and then checks the number of detections (step SB9) in the same manner as in the first stage of focus adjustment. When the number of times is detected, a comprehensive evaluation value is calculated (step SB10). Then, the focus changing unit 6 changes the focus setting value by a predetermined width (SB11). For example, when the minimum value of the focus setting value range is set in step SB1, a positive constant (hereinafter, this constant is referred to as a second focus setting movement width) so that the setting value gradually increases. To the value obtained by adding. In the second stage of focus adjustment, the focus setting value is finely adjusted for a character smaller than that in the first stage of focus adjustment. Therefore, the second focus setting movement width is smaller than the first focus setting movement width. To do. When the focus setting movement width is reduced, the adjustment time increases. However, since the range is narrowed down in advance in the first focus setting step, the second focus adjustment step can be completed in a relatively short time.

  If the focus setting value after the change is within the effective range, for example, if the focus setting value range is equal to or less than the maximum value of the focus setting value range, the process returns to step SB2 to repeat the process (step SB12). The second focus adjustment unit 8 specifies the focus setting value (step SB13), and the process of the second stage of focus adjustment ends. In step SB13, for example, a focus setting value that gives the maximum value of the comprehensive evaluation value obtained in step SB10 is selected.

  In this embodiment, the focus setting value that is optimal for a small character can be specified in a short time through the above steps. Further, by performing the adjustment process while switching the object for which the focus evaluation value is to be obtained, it is possible to specify a focus setting value that is in focus overall even when there are a plurality of types of objects to be detected and recognized. Although the object is a letter plate character, the present invention is not limited to this and may be applied to other objects. Moreover, although the example using the average value of edge strength and the contour waveform width has been described as the method for calculating the focus evaluation value, other methods may be used. Further, in the calculation of the comprehensive evaluation value in the first stage of the focus adjustment, the focus evaluation value is narrowed down. However, the same narrowing may be performed in the calculation of the comprehensive evaluation value in the second stage of the focus adjustment.

Embodiment 2. FIG.
A second embodiment of the present invention will be described below with reference to the drawings. The schematic configuration diagram is the same FIG. 1 as in the first embodiment, and the processing flow of this embodiment is substantially the same as that in the first embodiment, and is the same as FIG.

  In this embodiment, after the first focus adjustment unit 7 performs the first stage of focus adjustment (step S1), the second focus adjustment unit 8 performs the second stage of focus adjustment (step S2). Although the first stage of focus adjustment is the same as that of the first embodiment, the description thereof is omitted. Since the processing flow of the second stage of focus adjustment is substantially the same as that of the first embodiment, the processing of the second stage of focus adjustment will be described below with reference to the processing flowchart of FIG.

  First, in the same procedure as in the first embodiment, the second focus adjustment unit 8 initializes the focus setting value (step SB1), resets the counter / timer (step SB2), and the imaging unit 1 performs imaging (step SB3). ). Further, after the first detection / recognition process (step SB4) and the second detection / recognition process (step SB5) are performed, if the detection is successful (step SB6), the focus evaluation means 5 calculates the focus evaluation value. (Step SB7). In this embodiment, in this step SB7, the focus evaluation value is obtained after converting the image of the target area and reducing the influence of the reflected light. As the image conversion method, for example, the method described in JP-A-10-247239 is used. As a result of image conversion, it is possible to avoid calculating an incorrect focus evaluation value by regarding an edge caused by reflected light as a character edge. In this embodiment, the focus evaluation value is obtained not only for the second detection object but also for the serial number that is the first detection object.

  Thereafter, the second focus adjustment means 8 performs an elapsed time check (step SB8) and a detection frequency check (step SB9) in the same procedure as in the first embodiment, and a predetermined time has passed or a predetermined number of times have been detected. If so, a comprehensive evaluation value is calculated (step SB10). In this embodiment, the overall evaluation value is obtained using both the focus evaluation value obtained from the first detection object and the focus evaluation value obtained from the second detection object. As a calculation method of the comprehensive evaluation value, for example, an average value of the evaluation values of the second detection object obtained from an image in which the evaluation value of the first detection object is a predetermined value or more is used. As a result, when a foreign object or the like is stuck on the position of the second detection object, an error that calculates a comprehensive evaluation value larger than the actual value can be avoided.

  Then, the focus changing unit 6 changes the focus setting value by a predetermined width (SB11), and if the changed focus setting value is within the effective range, the process returns to step SB2 and repeats the process (step SB12). If not, the second focus adjustment means 8 specifies a focus setting value (step SB13), and the process of the second stage of focus adjustment ends.

  In this embodiment, through the above steps, proper focus adjustment can be performed even when reflected light is generated or a foreign object is attached to an object. In this embodiment, the image conversion for reducing the reflected light is performed at the time of calculating the focus evaluation value in the second stage of focus adjustment. However, this may be performed in the first stage of focus adjustment. Further, although the image conversion method is described in the literature, another method may be used. Further, as a calculation method of the comprehensive evaluation value, the average value of the focus evaluation values of the second detection object obtained from the image in which the focus evaluation value of the first detection object is a predetermined value or more is used. Another method may be used as long as focus evaluation is used in combination.

Embodiment 3 FIG.
A third embodiment of the present invention will be described below with reference to the drawings. FIG. 9 is a schematic configuration diagram. In the figure, reference numeral 25 denotes zoom evaluation means for obtaining a zoom evaluation value indicating how much the zoom of the image pickup means 1 matches a predetermined condition, and 26 changes the zoom setting value of the image pickup means 1 by a specified amount. The zoom changing means 27 is a zoom adjusting means for determining the zoom setting value from the zoom evaluation values in a plurality of input images picked up by sequentially changing the zoom setting value.
The processing flow of this embodiment is shown in FIG. Here, the operation will be described with reference to FIGS. FIG. 11 is a flowchart showing the processing procedure of the first stage of zoom adjustment / focus adjustment. FIG. 12 is a diagram for explaining a measurement position for obtaining a zoom evaluation value, and 28 is a measurement position.

  The processing flow in this embodiment will be first described with reference to FIG. In this embodiment, the zoom adjustment is added to the first stage of focus adjustment in the first embodiment (step T1), and then the second stage of focus adjustment (step T2) is performed. Since the processing of the second stage of focus adjustment is the same as that of the first embodiment, description thereof will be omitted, and the details of the first stage of zoom adjustment / focus adjustment of step T1 will be described below with reference to the flowchart of FIG.

  First, the zoom adjustment means 27 initializes the zoom value to a predetermined value (step TA1). For example, the smallest value that can be set. Thereafter, the focus value initialization (step TA2), the counter / timer reset (step TA3), and the image pickup by the image pickup means 1 (step TA4) are performed in the same procedure as steps SA1 to SA4 of the first embodiment. The first detection means 2 detects the first detection object, and the recognition means 4 performs recognition (step TA5). If the detection is successful (step TA6), the focus evaluation unit 5 calculates a focus evaluation value (step TA7), and the zoom evaluation unit calculates a zoom evaluation value (step TA8). As a method for calculating the zoom evaluation value, for example, the horizontal length 28 from the center position of the serial number at the left end to the center position of the serial number at the right end in FIG. 12 is used as an index. The difference is taken as the zoom evaluation value.

  Thereafter, as in steps SA7 to SA12 of the first embodiment, an elapsed time check (step TA9) and a detection count check (step TA10) are performed, and when a predetermined time has passed or a predetermined count has been detected, the comprehensive evaluation value Is calculated (step TA11), and the focus changing means 6 changes the focus setting value by a predetermined width (step TA12). If the changed focus setting value is not within the effective range (step TA13), the first focus adjustment is performed. The means 7 determines the focus setting value range (step STA14).

  Thereafter, the zoom changing means 25 changes the zoom setting value by a predetermined width (step TA15). For example, when the smallest value that can be set in step TA1 is set, the value is changed to a value obtained by adding a positive constant to the preset value so that the setting value gradually increases. If the zoom setting value after the change is within the effective range, for example, if it is less than the maximum value that can be set, the process returns to step TA2 to repeat the process (step TA16). The means 27 specifies the zoom setting value and outputs the corresponding focus setting value range (step TA17), and the processing of the first stage of zoom adjustment / focus adjustment ends. In the loop of step TA2 to step TA16, the focus setting value range is obtained for each zoom setting value. In step TA17, the zoom adjusting means 27 selects a focus setting value range corresponding to the zoom setting value at which the minimum zoom evaluation value (the zoom evaluation value closest to the target value) is obtained.

  In this embodiment, through the above steps, focus adjustment and zoom adjustment can be performed efficiently even from a state where both focus and zoom are out of focus.

  In the first to third embodiments, the detection / recognition method is described in the method described in “Mitsubishi Electric Technical Report Vol.62 No.2 pp.9-12“ Number Plate Recognition Technology ””. In addition, although an example of recognition after detecting an object has been described, the recognition process may not be performed when the recognition information is not used.

  The image processing apparatus according to the present invention is applicable to, for example, a vehicle number recognition apparatus that installs a camera on a road or the like and automatically recognizes a car license plate using an image captured by the camera.

BRIEF DESCRIPTION OF THE DRAWINGS It is a schematic block diagram which shows Embodiment 1 of this invention. It is a flowchart of the adjustment process in Embodiment 1 of this invention. It is a flowchart which shows the process sequence of a focus adjustment 1st step. It is a flowchart which shows the process sequence of a focus adjustment 2nd step. It is a figure which shows the display character on a license plate. It is a figure which shows the detection result of the 1st detection means. It is a figure which shows the result of a focus adjustment 1st step. It is a figure which shows the detection result of the 2nd detection means 3. It is a schematic block diagram which shows Embodiment 3 of this invention. It is a flowchart of the adjustment process in Embodiment 3 of this invention. It is a flowchart which shows the process sequence of a zoom adjustment and a focus adjustment 1st step. It is a figure explaining the measurement position for calculating | requiring a zoom evaluation value.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1; Imaging means, 2; 1st detection means, 3; 2nd detection means, 4; Recognition means, 5: Focus evaluation means, 6: Focus change means, 7; First focus adjustment means, 8; 2 focus adjustment means, 25; zoom evaluation means, 26; zoom change means, 27; zoom adjustment means.

Claims (7)

  An imaging unit that captures an image and uses it as an input image, a first detection unit that detects a first detection target from the input image, and a second smaller than the first detection target detected by the first detection unit A second detection means for detecting the detection object, a recognition means for recognizing the detection object detected by the first detection means and the second detection means, a first detection means and a second detection means. Focus evaluation means for obtaining a focus evaluation value indicating the degree of coincidence of the focus with respect to the detection target of the image pickup means from the detection result and the input image, and a first detection target in a plurality of input images picked up by successively changing the focus setting value The first focus adjustment unit that determines the range of the focus setting value from the focus evaluation value of the object, and the focus setting value is sequentially changed within the range determined by the first focus adjustment unit and images are captured A second focus adjustment unit that determines a focus setting value from focus evaluation values of second detection objects in a number of input images, and a focus change unit that changes a determined amount of the focus setting value of the imaging unit. Image processing device.   The second focus adjustment unit determines the focus setting value using both the focus evaluation value of the first detection object and the focus evaluation value of the second detection object. Image processing apparatus.   The focus evaluation unit performs a predetermined image conversion process on a region where a detection target exists in an input image, and obtains a focus evaluation value for the converted image region. Image processing apparatus.   The first focus adjustment unit and the second focus adjustment unit adjust the focus using the focus evaluation values of the first detection object and the second detection object whose positions in the input image are within a predetermined range. The image processing apparatus according to claim 1, wherein:   The first focus adjustment unit and the second focus adjustment unit focus using the focus evaluation values of the first detection object and the second detection object whose sizes in the input image are within a predetermined range. The image processing apparatus according to claim 1, wherein adjustment is performed.   The first focus adjustment unit and the second focus adjustment unit perform focus adjustment using a focus evaluation value of a detection object whose recognition result by the recognition unit matches a predetermined one. The image processing apparatus described.   Zoom evaluation means for obtaining a zoom evaluation value indicating how much the zoom of the imaging means matches a predetermined condition from the detection result of the first detection means, and a plurality of images obtained by sequentially changing the zoom setting value Zoom adjusting means for determining a zoom setting value from a zoom evaluation value in the input image, and zoom changing means for changing the zoom setting value of the imaging means by an amount determined by the zoom adjusting means. The image processing apparatus according to claim 1.

Images