JPH04344576A - Image processor - Google Patents

Abstract

PURPOSE:To automate monitoring of an object by specifically setting only the object in image data obtained by photographing the neighborhood of the object subject to monitoring to obtain color information of the object, and by identifying and outputting only the object in image data on the basis of the color information in measuring the object. CONSTITUTION:Video camera 13 is used to take a picture of the neighborhood of the position where oil leakage of actuator 12 that is an object is generated, and obtained image data is displayed on monitor television 15. Next, an area where only the object is present without taking picture of the background of screen is set by input pointing using a mouse or the like. Next, an image processing unit 14 is used to learn color information about the object. Further, actually, the neighborhood of actuator 12 is monitored by photographing using the camera 13, image data obtained using the image processing unit 14 is compared with color information of the object obtained when learning the color information of the object. Successively, each image element data in a measured area is displayed on the monitor television 15, and as a result a binary image is obtained where only the actuator 12 is extracted omitting other background in the measured area.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for identifying a target from an image.

0002

[Prior Art] Conventionally, in order to discover defective products and monitor discoloration on factory lines, etc., a colored liquid is sealed in a target object flowing through the line, and a supervisor can identify leaks.
I was trying to keep an eye on it.

0003

[Problem to be Solved by the Invention] However, in the above-mentioned monitoring work, it is necessary for a monitor to regularly stand by near the target and monitor it, and especially when monitoring for a long time or at night, the target object becomes difficult to monitor. There was a problem in that while the ability to identify objects deteriorated, the ability also varied depending on the individual differences of each observer.

The present invention has been made in view of the above-mentioned circumstances, and its purpose is to provide an image processing device that can automate the monitoring of a target without requiring a special observer. It is about providing.

[0005]

[Means and effects for solving the problems] That is, the present invention provides color information of the target, for example, 3 This method obtains the luminance distribution of primary color components, their average value, standard deviation, etc., and during measurement, only the target object in the image data is identified and output based on the above color information, and a trained monitor stationed near the target object is used. This makes it possible to greatly simplify the monitoring of targets without the need for personnel.

[0006]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows an example of an image processing apparatus applied to durability tests of aircraft hydraulic equipment and the like. In other words, it is used to monitor and measure whether oil leaks occur during operation when the equipment is operated for a long period of time, and if so, how much oil leaks. , 11 in the figure is a hydraulic drive source, 12 is this hydraulic drive source 1
This is an actuator, which is a hydraulic device driven by 1. A video camera 13 is installed and fixed so as to face the position where oil leakage of the actuator 12 is expected to occur, and the actuator 12 obtained by this video camera 13 is
Image data centered on is sent to the image processing device 14.

This image processing device 14 includes a video camera 1
By performing image processing with a target object identification function on the image data sent from 3, only the image of the actuator 12 that has caused the oil leak is selectively extracted. It is sent to the television 15 and output there. Next, the operation of the above embodiment will be explained.

[0009] At the beginning of operation, the target object is first learned in advance. This learning operation is performed when the actuator 1 becomes the target object.
This is done by using the video camera 13 to take an image of the area around the location where the oil leak may occur (No. 2). The image data obtained by imaging is directly displayed on the monitor television 15 via the image processing device 14, and once it is displayed,
For example, use a keyboard or mouse (not shown) as a setting device to input the coordinates of the upper left and lower right of the corresponding rectangular range by pointing to the area where only the target object exists without the background on the screen. Set by.

[0010] Once the area is set, the image processing device 14 executes learning according to the processing procedure shown in FIG. That is, first, the image processing device 14
The color image data sent from the computer that falls within the set learning area is digitized pixel by pixel using luminance values for each of the three primary color components of red, green, and blue (step S1). In this case, the value of each pixel data digitized by sequentially accessing the addresses within the area is treated as the luminance value of each of the three primary color components of light, as described above. For example, each of the primary color components of one pixel is
If data is processed as a total of 24 bits (256 gradations) and three primary colors, approximately 16.7 million colors can be expressed.

[0011] Once the image data in the setting area has been digitized, the luminance value distribution of each of the three primary color components is examined (step S2). Thereafter, the average value of each brightness value distribution obtained is determined (step S3), and the standard deviation is further determined based on this average value and the above brightness value distribution (step S3).
4). FIG. 4 illustrates the luminance value distribution, average value, and standard deviation of each of the three primary color components obtained in this way. Figure 4 (1) shows the luminance value distribution of the red component, where the average value is Er and the standard deviation is σ (not shown), the lower limit R1 (=Er-kσ, k is a constant), and the upper limit R2 (=
Er+kσ). Similarly, FIG. 4 (2) shows the lower limit G when the average value in the luminance value distribution of the green component is Eg.
1, the upper limit G2 is the lower limit B1, and FIG. 4 (3) is the lower limit B1 when the average value in the luminance value distribution of the blue component is Eg,
Indicates the upper limit value B2.

[0012] In this way, by obtaining the upper and lower limit values E±kσ of each color component from the average value E and standard deviation σ of the luminance values, a color filter (passing luminance range) for each color can be obtained (step S5), the obtained color information of the target object,
That is, a threshold value consisting of the upper and lower limits is set, and the processing related to target object learning is thus completed.

[0013] After learning the target object, the target object is actually measured next. This measurement operation is performed by again imaging and monitoring the area around the actuator 12, which is the target object, using the video camera 13, and the image data obtained by imaging is input to the image processing device 14 at any time. In the image processing device 14, in the same manner as in the setting of the learning area, the area around the actuator 12, which is the target object to be measured, is set by using, for example, a keyboard, a mouse, etc. (not shown) as a setting device, and setting the corresponding area. Set by pointing and inputting the coordinates of the upper left and lower right of the rectangular range. FIG. 5 shows the measurement area thus set and the learning area therein.

Once the measurement area is set, the image processing device 14 executes measurement for each pixel in the learning area according to the processing procedure shown in FIG. That is,
First, the image processing device 14 digitizes the color image data sent from the video camera 13 into luminance values for each of the three primary color components, red, green, and blue, for each pixel (step M1). In this case, the value of each pixel data digitized by sequentially accessing the addresses within the area is treated as the luminance value of each of the three primary color components, as in the above learning.

[0015] Then, each pixel data of the obtained digital value is compared with the color information of the target object obtained during the above-mentioned learning, and it is determined whether the pixel data is within the luminance value range indicated by the threshold value of the color information ( Step M2). In this case, as shown in FIG. 6, for each pixel data, for example, if the brightness value of the red component is present, it is determined whether or not it is within the brightness value range indicated by the threshold value of the color information, that is, the lower limit value R1 and the upper limit value R2. The judgment result is expressed as 1-bit data "1" or "0". Three 1s are determined for one pixel based on the judgment results of each of the three primary color components.
Since bit information is output, these are logically ANDed using an AND circuit. By displaying and outputting each pixel data in the measurement area obtained in this way on the monitor television 15 as shown in FIG. 7, other backgrounds in the measurement area are omitted and only the target object, the actuator 12, is extracted. It is obtained as a binarized image. Therefore, all pixel data within the outline of the actuator 12, which is the target object displayed on the monitor television 15, is "1",
If it is a uniform "solid display", it is normal, and if it is "0" partially and "uneven", it means that oil leakage has occurred, which can be easily identified.

[0016] In the above embodiment, an image processing device applied to a durability test of fixedly installed aircraft hydraulic equipment, etc. was shown, but as shown in FIG. It is also possible to individually judge whether the parts 22, 22, ... are normal based on their color information,
Identification can also be automated.

Furthermore, the output device for outputting the results obtained by the image processing device 14 is not limited to the monitor television 15, but may be connected to a printer, a measuring instrument, a computer, or the like.

[0018]

As described in detail above, according to the present invention, color information of the target, for example, 3 The luminance distribution of the primary color components, their average value, standard deviation, etc. are obtained, and during measurement, only the target object in the image data is identified and output based on the above color information, so it is easy for observers who are familiar with the vicinity of the target object. It is possible to provide an image processing apparatus that can greatly simplify the monitoring work since it is not necessary to have the image processing apparatus permanently resident and only the binarized extracted image of the target object needs to be monitored.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing an image processing apparatus applied to a durability test of hydraulic equipment, etc. according to an embodiment of the present invention.

FIG. 2 is a diagram showing a processing procedure when learning a target object.

FIG. 3 is a diagram showing a processing procedure when measuring a target object.

FIG. 4 is a diagram illustrating the luminance value distribution of each primary color component obtained in step S2 of FIG. 2;

FIG. 5 is a diagram showing a learning area in the measurement area displayed on the monitor television in FIG. 1;

FIG. 6 is a diagram showing the contents of the identification process executed in step M2 of FIG. 3 in units of pixels.

FIG. 7 is a diagram showing an image obtained as a result of measurement displayed on the monitor television in FIG. 1;

FIG. 8 is a diagram showing another application example of the image processing device according to the embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 11... Hydraulic drive source, 12... Actuator, 13... Video camera, 14... Image processing device, 15... Monitor television,
21...Line, 22...Parts.

Claims (1)

[Claims] 1. Imaging means for taking an image around a target to be monitored, first instruction setting means for specifying and setting the target in image data obtained by the imaging means, and the first instruction setting. A learning means for learning color information of the target object by the means, and an identification means for identifying and outputting only the target object in the image data obtained by the imaging means at the time of measurement based on the color information obtained by the learning means. An image processing device characterized by: