JP2949737B2 - Fruit detection method - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a method for detecting fruits and the like.

2. Description of the Related Art For example, in the case of an image of an apple tree, when taking out an apple tree which is a tree, the density, noise, and the like of the background of the apple tree and the leaves of the tree, soil, sky, etc. , And the relationship between the objects, etc., are factors for determining the image processing procedure.

SUMMARY OF THE INVENTION It is an object of the present invention to improve the detection accuracy when extracting an image of a fruit or the like from a background.

Means for Solving the Problems The present invention compares the magnitudes of green and blue signals among red, green and blue signal components of an object such as a red fruit input from a camera, and determines the larger one as a red signal. And a method for detecting a fruit or the like, characterized in that the fruit and its background color are separated by subtraction from the fruit.

Actions and Effects of the Invention A video of a red-based fruit or the like in front is captured by a camera, and among the input red, green, and blue signal components, components other than red, that is, the magnitudes of green and blue signals are compared, and the comparison is performed. By subtracting the larger one from the red light, fruits and the like are separated from the background color.

In this way, the result of subtracting the green and blue signals from the red signal,
Since the green and blue signals are minimal in the portion of the red fruits, the red signals of the red fruits are emphasized. As a result, the detection accuracy of red fruits and the like can be improved.

Embodiment In the illustrated example, an image of an object such as a fruit in front is captured by a camera 1 of an image sensor, and red R, green G, and blue B signals of each pixel are stored in an image frame memory 3 via an input interface 2. Is stored. These red, green,
The CPU receives the blue signal, calculates the color vector angle of each pixel, compares these color vectors with the white vector angle, and determines valid pixels and invalid pixels. ing. 4 is an image monitor, 5
Is an output interface.

Figures 2 and 3 show tomato leaves, stems,
This is an example of detection for detecting a red portion of a tomato fruit from other backgrounds. The image of the tomato is captured by the camera 1 and the red color of each pixel input from the camera 1 and stored in the image frame memory 5 is shown. , And the color vectors of the green and blue signals are calculated in the CPU, and the angle θ in the color vector direction with respect to the red signal is calculated. The white vector angle is α
Since 35.264 degrees, each pixel vector of θ ≧ α is extracted as an effective pixel, and the others are separated as an invalid pixel from the tomato fruit and the background.

FIGS. 4 and 5 show a configuration in which an image when the camera 1 is turned in the direction of the fruit is compared with an image when the camera 1 is turned to the background, and an object such as a fruit is detected based on the difference. And

For example, in a tomato detection example, a red signal of a background image is captured, and a red signal of a fruit direction image is captured. The difference between these two images is obtained, and a part where the difference satisfies the difference ≧ α is regarded as a fruit.

By separating the image portion of the fruit from the image captured by the camera 1 and the portion that is not, and extracting only the image portion of the fruit, it is possible to improve the detection accuracy of the fruit.

FIG. 6 and FIG. 7 show red, green,
Of the green light components, components other than the color of the object such as fruits, for example, in the case of tomatoes, compare the size of the green and blue lights, and subtract the larger one from the red light to separate the fruit and its background color Configuration.

For example, in detecting a tomato, the red, green and blue signals are fetched from the image frame memory 3 and the larger one of the green and blue signals is subtracted from the red signal. The fruit part is detected from this result.

In fruit detection in natural light, natural light is red,
Green and blue signals are almost equally included. Therefore, for example, in the detection of tomato, it is not possible to distinguish between the tomato and the background only with the red signal.

As a result of subtracting the green and blue signals from the red signal, the green and blue signals are minimal in the tomato portion, so the red signal is emphasized only in the tomato portion. By such a process, the detection accuracy of tomato can be improved.

[Brief description of the drawings]

1 shows an embodiment of the present invention. FIG. 1 is a control block diagram, FIG. 2 is a flowchart, FIG. 3 is an explanatory coordinate diagram, FIG. 4 is a flowchart of another detection method, and FIG.
FIG. 6 is an explanatory diagram, FIG. 6 is a flowchart of another detection method, and FIG. 7 is an explanatory diagram thereof. In the figure, reference numeral 1 denotes a camera, 2 denotes an input interface, 3 denotes an image frame memory, 4 denotes an image monitor, and 5 denotes an output interface.

Claims (1)

(57) [Claims] 1. A method for comparing the magnitude of green and blue signals among red, green and blue signal components of an object such as a red fruit input from a camera, and subtracting the larger one from the red signal to obtain a difference between the red and green signals. A method for detecting a fruit or the like, wherein the background color is separated.