JPS62279480A - Fruit recognizing device - Google Patents

Abstract

PURPOSE:To decrease as a whole the storage capacity of a storage means, and also, to efficiently execute a read-out processing by storing binarization image information being color information and variable density image information being brightness information, in a lump as one image data. CONSTITUTION:A red signal and a blue signal of color signals separated by an NTSC decoder 2 are brought to a subtraction processing by a color separating circuit 3, and also, binary-coded by a binarization circuit 4, and converted to binarization image information S4 which has extracted a specific color area corresponding to only a color of a fruit to be recognized. On the other hand, a luminance signal is quantized by an A/D converter 5 and converted to variable density image information S5, synthesized with the information S4 at every quantized picture element unit by a data buffer 6 and stored as image information of one screen portion. In such case, an image data S6 per one picture element is expressed by 8 bit length, and its lowest bit position and the remaining 7 bit portion are used as a value of the information S4 and a value of the information S5, respectively. Accordingly, binarization color information and brightness information corresponding to every quantized picture element shown by 8 bit length are stored as the data S6 in the same address position in the memory 7.

Description

Detailed Description of the Invention 3. Detailed Description of the Invention [Industrial Application Field] The present invention provides a method for guiding a manipulator by detecting the position of a fruit to be worked with respect to a fruit work machine, for example. Regarding a fruit recognition device used as a means for obtaining control information, specifically, by subtracting image information in a specific color range corresponding to the color of the fruit to be recognized and image information that does not include the specific color. , a color separation means for separating and extracting image information corresponding only to the color of the fruit to be recognized, a binarization means for binarizing the image information obtained by the color separation means based on a set threshold value, and each of the images. Did you image the same area as the information? quantization means for quantizing the gray level image information based on its gray level; The present invention relates to a fruit recognition device including a notation tα means for storing pixel positions in correspondence with each other.

[Conventional technology]

This type of fruit recognition device described above, for example, converts color image information of the fruit to be recognized into color image information, image information in a specific color range corresponding to the color of the fruit to be recognized, and image information that does not include the color of the fruit to be recognized. The color information is normalized with respect to brightness by subtraction, then binarized, and an area corresponding only to the color of the fruit to be recognized is extracted based on the binarized image information and the gray level of the captured image information. The two pieces of image information, including the quantized gray image information, are temporarily stored in a storage means such as an image memory so that the respective quantized pixel positions are at the same position, and based on the stored image information, Information corresponding to the position of the fruit to be recognized is obtained.

In order to store each pixel position of the binarized image information and the quantized gray image information in correspondence with each other, conventionally, the portions where there is no stored gray image information are processed as not being binary image information. Based on the principle that the binarized image information can be restored by doing this, by masking the gradation image information with the binarized image information, the gradation image information of the area corresponding only to the color of the fruit to be recognized is stored. (See, for example, Japanese Patent Application No. 128933/1982 and Japanese Patent Application No. 211585/1982, which were previously proposed by the present applicant).

[Problem that the invention seeks to solve]

However, the conventional configuration described above has the following disadvantages, and there is room for improvement.

Since the fruit to be recognized is generally spherical, the fruit surface in the captured image may have a highlight area that appears to shine white, which is different from the main fruit color, due to strong reflection. Since the brightness of this highlight area, which looks like a dirt floor on the screen, is saturated, it contains the specific color to be extracted and other color information at similar levels. When image information is subtracted from image information that does not include the specific color, the level is substantially the same, so it is greatly attenuated, and when the subtraction result is binarized, the specific color is Color information is missing. Therefore, in order to accurately extract a specific color range, it is necessary to perform a so-called fill-in process to complement the highlight area of the binarized image information.

However, in the conventional configuration described above, when masking the grayscale image information with the binary image information, if the binary image information is missing in the highlight area, the masked grayscale image information will also be missing. . Since there is no brightness information to determine whether the missing part of the image corresponds to a highlight area or image information of another color, for example, if the missing part is small, is determined to be a noise area and removed, and processing is performed such as replacing it with grayscale image information of the area that complements the value obtained by averaging the brightness of surrounding pixels. Therefore, there is a disadvantage that the result of the hole filling process becomes inaccurate.

Furthermore, it is necessary to repeatedly refer to the stored image information while performing the filling process, which not only complicates the processing but also slows down the processing speed.

Therefore, it may be possible to store both the binarized image information and the grayscale image information separately, but in that case, the storage capacity of the image information would increase, and the image processing would be performed while referring to two different image information. However, the disadvantage is that the processing becomes complicated and the processing speed becomes slow.

[Means for solving problems]

A characteristic configuration of the fruit recognition device according to the present invention is that an image data synthesizing means is provided for synthesizing the binary image information and the quantized grayscale image information as one image data for each quantized pixel; The structure is such that the combined image data is stored in the storage means, and its functions and effects are as follows.

[For production]

That is, since it is configured to store binarized image information, which is color information, and grayscale image information, which is brightness information, together in one image data, for example, in the above-mentioned hole-filling process, etc. A plurality of types of image data can be stored efficiently so that two pieces of image information, including color information and brightness information, can be read out simultaneously in - times of processing.

〔Effect of the invention〕

Therefore, the two image information, the binarized image information and the grayscale image information corresponding to the area of the specific color range, are combined into one image data and stored, so that the stored data is not lost. It has become possible to reduce the storage capacity of the storage means as a whole4 and to improve the efficiency of the reading process.

〔Example〕

Embodiments of the present invention will be described below based on the drawings.

Figure 1 shows color image information (S
o) is a block diagram of a fruit recognition device for obtaining positional information of a recognition target fruit, and is configured as a device for recognizing warm-colored fruits such as mandarin oranges and apples; As image quality information fla (S1) of a specific color corresponding to the color (F), a color signal (R) of the red component that constitutes the color image quality information a<SO) is separated and extracted, and an image that does not include the specific color is extracted. As information (S2), a blue component color signal (B) is separated and extracted, and the blue component color signal (B) is subtracted from the red component color signal (Fl) and binarized. Binarized image information (S4 ). Then, the brightness signal (Y) as grayscale image information corresponding only to the brightness information of the color image information (So) was quantized based on the gray level, and grayscale image information (S1) was separated and extracted. After that, the grayscale image information (S1)
is the center of gravity (p+) of the area masked with the binarized image information (S4) and the highlight area (Po) that appears bright.
By determining the position of the fruit, positional information that allows each of the fruits that appear to be overlapped to be separated and identified is obtained.

In other words, ↑Color video camera (1
) is output as color image information (see Figure 3 (a)).
So) is processed by the NTSC decoder (2) as synchronization signals of vertical synchronization signal (VD) and horizontal synchronization signal (+10), color signals of red signal (R) and blue signal ([1), and gray scale image information. luminance signals (Y).

The separated red signal (R) and blue signal (B) are image characteristics f of a specific color range corresponding to the recognition target fruit (F).
) 3 (S1) is subtracted by the color separation circuit (3) as a color separation means, and the binarization circuit (4)
The fruit to be recognized (F
) is converted into binarized image information (S4) (see FIG. 3 (rl)) in which a specific color area corresponding only to the color is extracted.

On the other hand, the luminance signal (Y) is processed by A/
In the D converter (5), the brightness, that is, the gray level, is quantized into 128 steps and converted into shading quality information (css) expressed as a 7-bit digital value, and the data is used as an image data synthesis means. In the buffer (6), every third quantized pixel is combined with the binarized image information (S4) and stored in the image memory (7) as an image storage means.
The image information is temporarily stored as image information for one screen. In the present embodiment, when the luminance signal ('/) is quantized, the screen light is converted into mosaic-like grayscale image information (S5) of 32 x 32 pixels (see Figure 3 (C) 1?'i ).

The synthesis of image data in which the binarized image information (S4) and the grayscale image information (S1) are synthesized as one image data (S1) for each quantized pixel will be explained as shown in FIG. - The image data (S6) per pixel is expressed as a whole with a length of 8 bits, and the lowest bit position (bit "0") is expressed as the binary image information (S4). It is used as a value, and the upper 7 bits are used as a value indicating the brightness of the grayscale image information (S5). However, above 2
When the value of the digitized image information (S4) is "1", it indicates a color region of the recognition target fruit (F), and "0" indicates a color region other than the recognition target fruit (F). be.

Therefore, the image memory (7) stores binarized color information and brightness information corresponding to each quantized pixel represented by 8-bit length as one image data (S6) at the same address position. I will do it.

However, the process of obtaining the binarized image information ID (s4) and i
A/D conversion of the above-mentioned variation signal Y) to digitize the gradation image characteristics? Process for obtaining H(S1) and the above 2
The process of combining and storing the digitized image information (S4) and the grayscale image information (S1) into one image data (S1) as image data in one-to-one correspondence for each pixel is performed using the vertical synchronization signal (VD) and the grayscale image information (S1). In the timing circuit (8), the timing circuit (8) performs simultaneous parallel operation in synchronization with the horizontal synchronization signal (I[D).
The processing operation is controlled. Furthermore, a strobe device (9) is activated in synchronization with the imaging operation by the camera (1).
) a strobe control circuit (10) activated by a control signal from the timing circuit (8) so as to emit light;
Accordingly, the grooves are formed so as to optimize the light emission timing and the light emission intensity.

Further, the binarized image information (S1) and the grayscale image information (S2) stored in the image memory (7) are converted into a composite video signal by a CRT controller (11) and a TV signal generator (12). , is displayed on a monitor television (13).

The binarized image information (S4) and the grayscale image information (S1)
), the control processor (CPU+) and the numerical calculation processor (CPUz) process the image data (sb) for each pixel.
Based on the image information stored as After performing the hole-filling process in Figure 3 (b)), the binarized image information (54,) that has been subjected to the hole-filling process (see Figure 3 (b)) is
Highlight area (PG), which is the peak brightness point, based on the gray level in the area where C) is “1”
(See Figure 3 (D)), and also the process of determining the position of the center of gravity (r'l) of the area where the binarized image information (S4) is "1", that is, the area corresponding to the fruit. administered.

To explain the hole filling process, for example, after performing a labeling process based on the value of the binarized image information tG (S1) of the synthesized image data (S6), Image information! By replacing the part where the value of 11J (S4-) is "0" with "1", the blanks related to the color information Q can be filled in without missing the brightness information of that part.

Next, the image information that has undergone the hole-filling process is relabeled, and, for example, only the maximum brightness value for each label and the area with a brightness one step lower than the maximum value are left, and the image data of other areas (S6 ) are all replaced with "0" to extract the highlight area (Po).

And the number of fruits recognized and their position information ta, that is, the highlight? An interface device (14) transmits information on the position coordinates of the iJl area (Po) and the center of gravity (Pl) to a host computer (CPt1.) that uses the recognition results.
In addition to transmitting information through the system, the entire operation is controlled. In FIG. 1, (15) represents each of the processors (
CP[I2]), (CP[I2) operating programs and calculation data, and the host computer (CPI2)
This is a memory for storing various data exchanged with the IQ).

Next, a description will be given of means for guiding the working cocoon break of the fruit working machine using the positional information of the highlight area (Po) and the center of gravity (p+) obtained by the fruit recognition device configured as described above.

As shown in FIG. 4, a boom (17) is pivotally supported on a traveling body (V) equipped with a pair of left and right traveling wheels (16) at the front and rear so as to be able to rise and fall and swing from side to side.
), an auxiliary boom (18) is pivotably supported horizontally, and a multi-articulated telescoping eyebrow break (19) is attached to the tip of the auxiliary boom (18) in vertical and horizontal directions. It is attached so that it can swing freely.

Then, a working hand (11) is attached to the tip of the manipulator (19), and the manipulator (19) is placed in a state where the viewing direction matches the direction of the manipulator (19), that is, the working hand (H). The canorak 1) is housed inside the base end of the canorak 1).

Therefore, from the camera (1), the eyebrow break (
19) Image information (
So) is obtained, and by performing the above-mentioned processing on the female image information f[1i (S1), a highlight corresponding to the position of the work target fruit (F) with respect to the current position of the manipulator (19) is obtained. Information on the position coordinates of the area (Po) and center of gravity (Pl) can be obtained.

Then, the cocoon break (19) is protruded in the direction of the center of gravity (r'+), and based on the detection information from an infrared reflective proximity sensor, etc., the fruit (F) to be worked is However, as the manipulator (19) approaches the fruit (F) within a set distance, the direction of the manipulator (19) is corrected in the direction of the highlight area (PG), and then the fruit (F) approaches the fruit (F). It is possible to accurately approach a predetermined one of the fruits (P) recognized as a region.

[Another example]

In the above embodiment, both the binarized image information +D (S1) and the grayscale image information (S1) are used to recognize the target fruit (F
) The highlight area (Po) and the center of gravity (Pl) of the above-mentioned highlight area (Po) are obtained separately, but the present invention is applicable to the highlight area (Po) and the center of gravity (Pl) of Of course, it can also be applied to the case where only one of the iJT area (Pa) or the center of gravity (Pl) is determined.
Further, the processing of the stored image information can be changed in various ways depending on the purpose of the device to which it is applied.

Furthermore, in the above embodiment, in order to obtain image information corresponding only to the color of the fruit to be recognized (F), the red and blue color signals constituting the color image are directly subtracted. The image information of the color to be subtracted may be set in accordance with the color of the target fruit, and various changes can be made depending on the color of the recognition target fruit. Furthermore, the number of pixels to be quantized may be set according to the resolution required for the information to be recognized, and the present invention is not limited to the above embodiments.

[Brief explanation of the drawing]

The drawings show an embodiment of the fruit recognition device according to the present invention.
The figure is a block diagram showing the configuration of the fruit recognition device.
FIG. 3 is an explanatory diagram showing the structure of stored image data, FIG. 3 is an explanatory diagram of image information, and FIG. 4 is a side view showing the schematic structure of the fruit working machine. (F)... Recognition target fruit, (Sl)...
・Image information in a specific color range, (S2)...Image information that does not include a specific color, (SJ)...Image information that corresponds only to the color of the fruit to be recognized, (S4)・・・・・・・
Binarized image information, (Y)... Grayscale image information, (
S1)...Quantized grayscale image information, (S6
)... Image data, (3)... Color separation means, (4)... Binarization means, (5)...
... Quantization means, (6) ... Image data synthesis means, (7) ... Storage means.

Claims (1)

[Claims] Image information (S_1) of a specific color range corresponding to the color of the recognition target fruit (F), and image information (S_1) that does not include the specific color.
2), the recognition target fruit (F) is obtained by subtracting
a color separation means (3) for separating and extracting image information (S_3) corresponding only to the color of , and a binary value for binarizing the image information (S_3) obtained by the color separation means (3) based on a set threshold value. converting means (4), each of the image information (S_1) to (S
Quantization means (5) for quantizing the grayscale image information (Y) captured in the same range as _3) based on its gray level;
and the quantized grayscale image information (S_5) and the binarized image information (S_5) binarized by the binarization means (4).
A fruit recognition device comprising a storage means (7) for storing the binarized image information (S_4) and the quantized gray image in a state in which the respective quantized pixel positions correspond to each other. image data synthesis means (which synthesizes the information (S_5) as one image data (S_6) for each quantized pixel;
6), and the synthesized image data (S_6) is
A fruit recognition device configured to be stored in the storage means (7).