JPH08266153A - Device for detecting crop - Google Patents

Abstract

PURPOSE: To provide a crop-detecting device capable of rapidly evaluating the growth states of crops, while avoiding the complication of the device, when the plural crops as evaluation targets are planted in an arranged state. CONSTITUTION: This crop detection device is provided with a color type image pick-up means S for imaging a planted surface on which plural crops N are planted in an arranged state, a crop region-extracting means for extracting a crop region corresponding to the plural crops N by the color of the crops N based on the image pick-up information from the means S, and an evaluation means for evaluating the respective growth states of the crops N on the planted surface on the basis of the crop region corresponding to the crops N.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to image pickup means for picking up a planting surface on which a crop is planted, and crop area extracting means for extracting a crop area corresponding to the crop based on the image pickup information of the image pickup means. , A crop detection device provided with an evaluation means for evaluating the growth state of the crop on the planting surface based on the information of the crop area extraction means.

[0002]

2. Description of the Related Art In the above crop detecting device, conventionally, a planting surface on which vegetables and flowers as crops are planted is colored by a CCD.
An image is picked up by a color type image pickup means such as a camera, and based on the image pickup information, for example, pixel information at a predetermined resolution in the vertical and horizontal directions of the image pickup screen, the crop region corresponding to the above vegetables and flowers is used as a background region such as soil. From the different color information
Based on the crop area information, for example, the growth state such as whether or not the growth of the crop is evaluated (for example, Table 3-50
4424).

[0003]

However, in the above-mentioned conventional technique, since the crop area is extracted and the growth state is evaluated based on the information obtained by individually imaging the crop,
When the crops to be evaluated are, for example, a plurality of seedlings planted in an aligned state, the crops or the imaging means are moved so that each seedling is located within the imaging visual field, and at each position. It is necessary to sequentially evaluate each seedling based on the imaged information. Therefore, there is a problem that it takes a long time to evaluate all of the plurality of seedlings, and the device becomes a complicated device having a moving mechanism for the crop or the imaging means.

The present invention has been made in view of the above circumstances, and an object thereof is to avoid complication of the apparatus even when a plurality of crops to be evaluated are planted in an aligned state. The purpose is to quickly evaluate the growth condition of each crop and eliminate the above-mentioned drawbacks of the prior art.

[0005]

A first characteristic configuration of a crop detecting apparatus according to the present invention is characterized in that a plurality of the crops are planted in an aligned state on the planting surface, and the crop area extracting means is arranged to produce the crop area. Is extracted as a crop area corresponding to each of the plurality of crops, and the evaluation unit evaluates the growth state of each of the plurality of crops based on the information of each crop area corresponding to each of the plurality of crops. It is configured as follows.

A second characteristic configuration is the same as the first characteristic configuration, wherein the image pickup means is a color type image pickup means, and the crop area extraction means extracts the crop area from the crop. It is configured to be done by the color of.

A third characteristic configuration is the first or second aspect described above.
In the characteristic configuration of the above, noise correction means for correcting dark level noise in the image pickup information of the image pickup means is provided,
The crop area extracting unit is configured to extract the crop area based on the image pickup information of the image pickup unit whose dark level noise has been corrected by the noise correction unit.

A fourth characteristic configuration is that the plurality of crops are planted on a tray for transportation in the first, second or third characteristic configuration.

[0009]

According to the first characteristic configuration of the crop detecting apparatus according to the present invention, an image of a planting surface on which a plurality of crops are planted in an aligned state is imaged, and a plurality of images are obtained from imaging information including the plurality of crops. The crop areas corresponding to each of the crops are extracted,
From the crop area information of each of the plurality of crops, the growth state of each of the plurality of crops on the planting surface is evaluated.

According to the second characteristic constitution, in the first characteristic constitution, the planting surface on which a plurality of crops are planted in an aligned state is imaged by the color type image pickup means,
Based on the color imaging information including the plurality of crops, the crop regions of each of the plurality of crops are extracted by the color of the crop, and the growth state of each of the plurality of crops is evaluated.

According to the third characteristic constitution, in the first or second characteristic constitution, the image pickup information including a plurality of crops in which the dark level noise is corrected and the influence of the dark level noise is removed. Based on the above, the crop regions of each of the plurality of crops are extracted, and the growth state of each of the plurality of crops is evaluated.

According to the fourth characteristic configuration, the first,
In the second or third characteristic configuration, a transport tray on which a plurality of crops are planted in an aligned state is transported to an image capturing position of an image capturing unit, an image of a planting surface on the tray is captured, and the plurality of crops are captured. Based on the imaging information of the planting surface on the tray including, the crop regions of each of the plurality of crops are extracted, and the growth state of each of the plurality of crops is evaluated.

[0013]

Therefore, according to the first characteristic configuration of the crop detecting apparatus according to the present invention, a plurality of crops planted in an aligned state are imaged at once, and imaging information including the plurality of crops. Is processed to evaluate the growth state of each of the plurality of crops at the same time, the crops or the imaging means are moved so that each crop is located at the imaging position, and the imaging information at each position is recorded as in the conventional method. It is not necessary to sequentially evaluate each crop on the basis of the above, and therefore, it is possible to quickly evaluate the growth state of each crop while avoiding the complication of the apparatus because the moving mechanism for the crop or the image pickup means is unnecessary. Came to be able to do.

Further, according to the second characteristic configuration, since the crop area is extracted by the color of the crop based on the color image information, for example, when the crop area is extracted based on the monochrome image information of only the brightness information. In contrast, when the crop and the background have different colors but the brightness is equal, it is not possible to extract the crop area accurately, whereas the color and the background can be distinguished to extract the crop area accurately. As a result, suitable means for implementing the first characteristic configuration can be obtained.

Further, according to the third characteristic configuration, since the crop region is extracted based on the imaging information from which the dark level noise is removed, for example, the brightness and color of the crop can be determined based on the imaging information including the dark level noise. When the crop area is extracted by the above method, there is no inconvenience that the brightness or color of the crop is deviated from the original brightness or color and the crop area cannot be properly extracted. Can be extracted, and a suitable means for implementing the first or second characteristic configuration can be obtained.

Further, according to the fourth characteristic configuration, since the transport tray on which a plurality of crops are planted is moved to the image pickup position and its image pickup and growth state are evaluated, for example, on the transport tray. Instead of imaging a plurality of crops planted directly on the ground and evaluating their growth state, it is necessary to accurately configure a moving mechanism for moving the imaging means to the position of the crops with little vibration. Although the device is complicated due to its nature, it can be carried out by a simpler moving mechanism for moving the tray or by manual transfer, and therefore, when carrying out the first, second or third characteristic configuration described above. Suitable means of

[0017]

EXAMPLES Examples of the present invention will be described below with reference to the drawings when applied to a seedling raising plant for growing cellular seedlings such as vegetables and flowers as crops.

As shown in FIG. 1, in a seedling raising plant, a plurality of cell-like seedlings N such as vegetables and flowers are planted on a rectangular-shaped seedling raising tray 1 for transportation in an aligned state in the vertical and horizontal directions. There is. A microcomputer provided with a color CCD camera S as a color image pickup means for picking up the planting surface (upper surface) of the seedling raising tray 1, an image processing section 3 and a control section 4 for controlling the entire apparatus. On the lower side of the moving body 2 in which the controller H used and the supplementary planting mechanism 5 and the like are mounted and suspended on the rail 16, the seedling raising tray 1 is provided.
Are conveyed by the belt conveyor 15 in the order of the crop recognition position (A) and the defective seedling supplementation position (B).

As shown in FIG. 2A, the CCD camera S takes an image of the planting surface of the seedling raising tray 1 located at the crop recognition position (A) from directly above, and the image pickup information is sent to the controller H. Entered and its controller H
The growing condition of each seedling N, such as the presence or absence of seedlings, is recognized inside. Then, in accordance with the recognition result, at the defective seedling supplementary planting position (B), the controller H is configured to operate the supplementary planting mechanism 5 so as to take measures such as replanting the defective seedling into a good seedling. There is. Although not shown, the moving body 2
A moving mechanism including an electric motor or the like for moving the carriage along the rail 16 is provided, and the moving mechanism is operation-controlled by the controller H.

In the supplementary planting mechanism 5, a movable arm 6 having a holding portion 6a for retaining the removed defective seedlings or seedlings for replanting is provided on the tip side, and the moving body 2 is located at the supplementary seedling position. A drive mechanism 5A is provided for moving the movable arm 6 toward and away from a predetermined seedling position in the moved state. The operation of the drive mechanism 5A is controlled by the controller H.

Next, the structure of the image processing unit 3 will be described. As shown in FIG. 3, R from the CCD camera S,
The video signals of G and B colors are first input to the video signal processing and A / D conversion circuit 30, and for each pixel of a predetermined resolution in the vertical and horizontal directions of the screen, for example, 8-bit (256 levels) digital colors. After being converted into image information, it is stored in the frame buffer M for each color. The dark level noise of each color image information stored in the frame buffer M is corrected by the dark noise correcting circuit 32 as a noise correcting means for correcting the dark level noise in the image pickup information of the CCD camera S, and the dark level noise is corrected for each pixel. The corrected color image information R, G, B is converted by the chromaticity conversion circuit 33 into normalized color information, that is, chromaticity information r, g, b. Then, according to whether or not the chromaticity information r, g, b of each pixel is within the preset chromaticity range of the seedling N in the color extraction circuit 34, as shown in FIG. The crop area corresponding to N, that is, the seedling area Tn is extracted, and the seedling area extraction information is stored in the image memory 36. An aperture control circuit 31 that controls the aperture of the CCD camera S, an image processor and memory controller 35 that exclusively executes image processing such as dark level noise correction, chromaticity conversion, and color extraction, and an image at each processing. And an image output circuit 37 for displaying on a CRT monitor or the like.

The control unit 4, as shown in FIG.
U4A and a drive control circuit 4B that outputs a drive signal to the supplementary planting mechanism 5 and the like. The image processing circuits M, 32, 33, 34, 35, 36 and 37 are connected to each other via an image data bus. Also, CP
The above-mentioned circuits M, 30,
31, 32, 33, 34, 35, 36, 37, etc., and the drive control circuit 4B are connected.

From the above, using the image processing unit 3,
Based on the image pickup information of the CCD camera S, the crop area extracting means 100 for extracting the crop area (seed area Tn) corresponding to the seedling N by the color of the seedling N is corrected by the dark noise correction circuit 32 for dark level noise. The seedling area Tn corresponding to the seedling N is extracted based on the captured image information of the CCD camera S. Further, using the image processing unit 3 and the CPU 4A, the evaluation unit 101 for evaluating the growth state of the seedling N on the planting surface on the tray 1 is configured based on the information of the crop region extraction unit 100. Then, in response to the plurality of seedlings N being planted, the crop area extracting unit 100 extracts the seedling areas as seedling areas Tn corresponding to each of the plurality of seedlings N, and the evaluating unit 101: The growing state of each of the plurality of seedlings N is evaluated based on the information of each seedling region Tn corresponding to each of the plurality of seedlings N.

Next, a concrete configuration of each circuit of the image processing unit 3 and the control unit 4 will be sequentially described. In the video signal processing and A / D conversion circuit 30, as shown in FIG. 4, the image pickup signal from the CCD camera S is first gain-adjusted by the video signal processing unit 7 so as to have a predetermined signal level. , R, G, B color signals and synchronization signal SYNC (1
/ 30 seconds sync signal) and output. The synchronization signal SYNC is supplied to each section as a timing signal, and the video signal processing section 7 sends an electronic signal to the CCD camera S based on a shutter speed change command from a switching control section 12 described later. A control signal for changing the shutter speed of the shutter in synchronization with the synchronization signal is output. The R, G, and B color signals are quantized for each pixel in each A / D conversion circuit 8 while being compared with a predetermined reference voltage f, and converted into 8-bit digital image information. And each 8
The bit color information is input to the R, G, and B frame buffers M that store the digital image information in pixel units via the switching unit 9, respectively.

The CCD cameras S have the same object (1
The planting surface of one seedling raising tray 1) is repeatedly imaged a predetermined number of times, and each color signal of R, G, B is A / D for each imaging.
It is configured to output to the conversion circuit 8. And
The switching is performed so that the value of each pixel of the digital image information (8-bit color information) output from each A / D conversion circuit 8 is changed from a small value to a large value each time the CCD camera S repeatedly captures an image. The control unit 12 is configured. That is, the switching control unit 12 gives the CCD camera S a shutter speed change command for changing the electronic shutter speed from a fast condition to a gradually slower condition a predetermined number of times. As a result, the exposure state of the CCD camera S in the repeated imaging is gradually changed from a dark condition to a bright condition, and the value of the pixel is changed from a small value to a large value.

Further, the maximum value of the respective amplitudes of the 8-bit color information is detected by the maximum value detection circuit 10, and the maximum value Y of the color information is set by the comparator 11 for determining the saturated state. The value X is compared. Here, the set value X for determining the saturation state is set to a value that is as large as possible (for example, about 250 levels in the case of 8 bits) at a level that does not saturate. The comparison output of the comparator 11 is input to the switching control unit 12 for controlling the switching state of each switching unit 9 to switch the information stored in each frame buffer M. That is, the maximum value Y of the color information is the set value X
If it does not exceed (X ≧ Y), the value of each pixel of the digital image information output from each A / D conversion circuit 8 is determined to be appropriate, and each color information after the A / D conversion is determined. It is stored in each frame buffer M. On the other hand, when the maximum value Y of the color information is larger than the set value X, the A / D
It is determined that the value of each pixel after conversion is not appropriate, and each color information already stored and held in each frame buffer M is held as it is.

Next, the dark noise correction circuit 32 will be described. The image pickup information at the time of darkness when the aperture of the CCD camera S is fully closed is fetched a plurality of times (16 times, etc.), and the dark level noise Vd for each pixel is obtained for each color R, G, B by averaging the number of times. As shown in FIG. 5, they are stored in the dark noise image memory 32a for each color. Then, as shown in FIG. 6, the correction calculation circuit 32b for each color is configured by a subtraction circuit that subtracts the dark level noise Vd from the output signal Vm before correction and outputs the output signal Vo after correction. There is. Therefore, the influence of the dark level noise Vd becomes larger as the lightness I of the imaging target is smaller, that is, the output signal V of the CCD is smaller. In the figure, Imax is the maximum value of the brightness I, and Vmax is the maximum value of the output signal V of the CCD.

[0028]

## EQU1 ## Vo = Vm-Vd

Next, the chromaticity conversion circuit 33 will be described. As shown in FIG. 7, the sum (R + G + B) of the image information of three colors is added to the chromaticity LUT (look-up table) 33b for each color configured to perform the conversion operation represented by the following equation.
And the respective color information R, G, B are input, and the respective chromaticity information r, g, b is output from each chromaticity LUT.

[0030]

## EQU2 ## r = R * 255 / (R + G + B) g = G * 255 / (R + G + B) b = B * 255 / (R + G + B)

Next, the color extraction circuit 34 will be described. As shown in FIG. 8, in the color range LUT (look-up table) 34a in which the information on the color range of the seedling N is set, the r chromaticity and the g chromaticity of the obtained chromaticity information r, g, b. When the information is input and the color coordinates indicated by the r and g chromaticities are within the color range of the seedling N, 1 is output as the extracted pixel data, and when it is not within the color range of the seedling N, extraction is performed. 0 is output as pixel data.

Here, the color range of the seedling N is represented by a range Cn of a predetermined shape in a two-dimensional coordinate space in which the g chromaticity is the horizontal axis and the r chromaticity is the vertical axis, as shown in FIG. It The range Cn is created by the procedure shown in the flowchart of FIG. First, reference images under various conditions such as weather conditions such as sunny and cloudy conditions and changing seedling samples are input, and each color information is converted into chromaticity r and g information. Then, on the monitor screen displaying the chromaticity r, g converted image, a predetermined seedling portion is designated with a light pen or the like, and the average value and standard deviation σ of the chromaticities r, g on the screen are obtained. ,
A rectangular range Cn indicated by a range of 3 times the standard deviation σ on both sides centering on the average value along the axis of each r chromaticity and g chromaticity
1, Cn2, etc. are obtained. Here, Cn1, Cn2, etc. are
The range obtained from the reference image of each of the different conditions above,
The range obtained by adding these is the range Cn.

Next, the control unit 4 and the image processing unit 3
The control operation will be described with reference to the flowcharts of FIGS. In the entire flow (FIG. 10), the seedling raising tray 1 to be inspected on the belt conveyor 15 is CCD.
After moving to the image pickup position of the camera S and performing the process of inputting image information about the planting surface of the inspection tray 1, defective seedlings (including seedling-less seedlings are included) based on the color image information obtained by the image input. ) Image processing to detect
Next, a supplemental planting process is performed in which the cellular seedlings at the location where the defective seedling is detected are removed, and a substitute cellular seedling is supplemented at the removed location.

In the image input (FIG. 11), first, CC
The D camera S performs an image pickup operation under the condition that the exposure state thereof is the darkest (the electronic shutter speed is the fastest), and the image pickup signals of the respective colors R, G, B are A / D converted for each pixel. The maximum value of the digital image information of each color after the A / D conversion is detected. The digital image information of each color is stored in each frame buffer M only when the maximum value does not exceed the set value for determining the saturated state.
Incidentally, since the first exposure state is the darkest condition, each frame buffer M always stores the digital image information of each color by the first image pickup. Next, the CCD camera S is gradually increased in brightness so as to change the stored information to a better pixel value in a brighter exposure state within the limit of not being saturated, while the shutter speed is decreased by the set amount. The flow of the storage processing from the image pickup operation to the frame buffer M is repeated for all the exposure conditions of a predetermined number of times.

In the image processing (FIG. 12), first, correction for removing dark level noise is performed, then smoothing processing is performed, and then chromaticity image conversion and color extraction are performed to perform binary conversion. Get seedling area information. Then, in the seedling area information, reduction processing and expansion processing are performed in order to remove the noise image. Next, since the tray 1 is set to a predetermined position with respect to the field of view of the CCD camera S, as shown in FIG.
As shown in (b), a grid-shaped mask MK is set on the image pickup screen of the CCD camera S and divided into screen portions corresponding to the positions of the respective seedlings N. Defective seedling detection processing is performed to detect the presence of missing or defective seedlings.

In the defective seedling detection process (FIG. 13), the area of each seedling area Tn is calculated, and the average area of the seedling area Tn is calculated excluding the missing strains. Then, each seedling parameter is calculated as an area ratio of each seedling region Tn to the average area, and the seedling parameter of each seedling N is compared with a preset defective seedling parameter. A seedling N having a seedling parameter larger than the bad seedling parameter is a normal seedling, but a seedling having a seedling parameter smaller than the bad seedling parameter is determined to be defective, and the above determination is performed for all seedlings N on the screen.

The process of setting the defective seedling parameters will be described. As shown in FIG. 14, the image on the tray 1 containing the defective seedlings is input as a reference image, and each seedling region Tn is extracted by color extraction. And then remove seedling area T
Calculate the average area of n. Then, the defective seedling is designated while looking at the monitor screen, and the defective seedling parameter is set as the area ratio of the seedling area Tn of the defective seedling to the average area.

[Other Embodiments] Next, another configuration for designating the color range of the seedling N for extracting the seedling area will be described with reference to FIGS. 16 and 17.
It will be described based on. Similar to the present embodiment (see FIG. 15), the reference image of various different conditions is input according to the procedure shown in the flowchart of FIG. 16, and each color information is converted into chromaticity r and g information. Then, on the monitor screen displaying the chromaticity r, g converted image, when a range is designated so as to surround a predetermined seedling portion with a light pen or the like, the pixels within the range are automatically divided into the seedling portion and the background portion. Then, as shown in FIG. 17, the chromaticity range Cn of a predetermined shape corresponding to the chromaticity r and g coordinates of the seedling portion becomes the color range of the seedling N. This range C
Similarly to the present embodiment, n is also set as a range obtained by adding the ranges obtained from the reference images under the different conditions described above.

In the above embodiment, the image pickup means is constituted by the color type image pickup means (color CCD camera S), but not the color type image pickup means, but the black and white type image pickup means (black and white type CC).
D camera etc.). Also, when the image pickup means is constituted by a color type image pickup means, the color CCD camera S
Instead, various image pickup means such as an electron tube type color television camera can be used.

In the above embodiment, the crop area extracting means 100 is used.
To extract the crop area, the three primary color information R, G, B output from the color type image pickup means is converted into chromaticity information r, g, b, and based on the chromaticity information r, g, b Although the crop area is extracted based on the color of the crop, the invention is not limited to this. For example, the crop area may be extracted based on the brightness of the crop based on black-and-white imaging information (brightness information).
Also, when the crop area is extracted by color based on the color image information, the three primary color information R, G, B are converted into the chromaticity information r,
The crop region may be extracted based on the three primary color information R, G, and B without being converted into g and b, or may be performed based on other color information.

In the above embodiments, the crop detecting apparatus of the present invention was described for the case of evaluating the growing condition of seedlings such as vegetables and flowers as crops, but the present invention is not limited to this, and seedlings of vegetables and flowers are not limited to this. It can be applied to devices for growing various crops other than. Further, the evaluation of the growth state is not limited to the one in which the poor growth including the defective strain is determined as in the above-mentioned Examples, and various growth states can be evaluated.

In the above embodiments, crops (seedlings of vegetables, flowers, etc.)
In order to plant a plurality of plants in a row on the planting surface in an aligned state, they are arranged in a grid pattern in the vertical and horizontal directions, but the present invention is not limited to this, and they may be planted in a honeycomb-shaped array state, for example.

In the above embodiment, a plurality of crops (seedlings such as vegetables and flowers) are planted on the tray 1 having a rectangular shape for transportation, but the shape of the tray 1 is not limited to the rectangular shape.
Further, it is not always necessary to plant it on the tray 1, and it may be planted directly on the ground or the like. In this case, the moving body 2 equipped with the crop detection device moves in a wide range to evaluate the growing condition of the crop.

It should be noted that reference numerals are given in the claims for convenience of comparison with the drawings, but the present invention is not limited to the configurations of the accompanying drawings by the entry.

[Brief description of drawings]

FIG. 1 is a partial side view showing a seedling raising plant equipped with a crop detection device.

FIG. 2 is a diagram showing a picked-up image of a crop on a tray and its area extraction.

FIG. 3 is a schematic block diagram showing a control configuration of the crop detection device.

FIG. 4 is a circuit block diagram showing a configuration of an image processing unit.

FIG. 5 is a block diagram of a dark level noise correction circuit.

FIG. 6 is a graph showing dark level noise correction characteristics.

FIG. 7 is a block diagram of a chromaticity conversion circuit that converts color image information into chromaticity image information.

FIG. 8 is a block diagram of a color extraction circuit that extracts a seedling area from chromaticity image information.

FIG. 9 is a chromaticity diagram showing a chromaticity range for seedling area extraction.

FIG. 10 is a flowchart of the overall control operation.

FIG. 11 is a flowchart of image input.

FIG. 12 is a flowchart of image processing.

FIG. 13: Flow chart for defective seedling determination

FIG. 14: Flow chart for setting defective seedling parameters

FIG. 15 is a flowchart of chromaticity range setting for seedling area extraction.

FIG. 16 is a flowchart of setting a chromaticity range for seedling area extraction according to another embodiment.

FIG. 17 is a chromaticity diagram showing a chromaticity range for seedling area extraction according to another embodiment.

[Explanation of symbols]

 N crop S imaging means 100 crop area extraction means 101 evaluation means 32 noise correction means 1 tray

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Shoji Tanaka 64, Ishizukita-machi, Sakai City, Osaka Prefecture Kubota Sakai Factory Co., Ltd.

Claims (4)

[Claims] 1. An imaging means (S) for imaging a planting surface on which a crop (N) is planted, and a crop region corresponding to the crop (N) is extracted based on imaging information of the imaging means (S). The crop area extracting means (100) for performing the crop area extracting means (100) and the evaluating means (101) for evaluating the growth state of the crop (N) on the planting surface based on the information of the crop area extracting means (100) are provided. A crop detection device, wherein a plurality of the crops (N) are planted in an aligned state on the planting surface, and the crop area extracting means (100) corresponds the crop areas to each of the plurality of crops (N). And the evaluation means (101) evaluates the growth state of each of the plurality of crops (N) based on the information of each crop area corresponding to each of the plurality of crops (N). Is configured to do Crop detection device. 2. The image pickup means (S) is constituted by a color type image pickup means (S), and the crop area extracting means (10).
0) is the crop detection device according to claim 1, wherein the crop area is extracted by the color of the crop (N). 3. A noise correction means (32) for correcting dark level noise in the image pickup information of the image pickup means (S) is provided, and the crop area extraction means (100) is provided by the noise correction means (32). The crop detection device according to claim 1 or 2, which is configured to extract the crop region based on imaging information of the imaging unit (S) in which dark level noise is corrected. 4. The crop detection device according to claim 1, 2 or 3, wherein the plurality of crops (N) are planted on a tray (1) for transportation.