White Hat Detection Device

Abstract

The present disclosure relates to white cap detection equipment. An apparatus (100) for analyzing a grain sample includes a light source (130, 140), an image sensor (120), and a controller (150). The light source (130, 140) is configured to illuminate the grain sample. The image sensor (120) is used to capture an image of the grain sample. The controller (150) is coupled to the image sensor (120) to receive the image of the grain sample therefrom and is configured to analyze the image to detect at least one material other than the grain in the grain sample. The light source (140) is configured to illuminate the grain sample with a local light spot having a size smaller than the average wheat grain width. Image analysis and detection of materials other than the grain can be performed using at least in part a trained neural network and other artificial intelligence (AI) algorithms.

Description

Technical Field

[0001] The present invention relates to an apparatus for analyzing a grain sample and a combine harvester having such an apparatus. The present invention also relates to a method of analyzing a grain sample using the apparatus. Background Art

[0002] A combine harvester is a complex agricultural machine that travels through a field to harvest a grain crop while separating the grain ears from the plants and the grains from the ears. In some grain crop varieties, such as wheat, the grains are enclosed in husks that are separated from the grains in a threshing, separating, and cleaning unit. The separated grains (also referred to as clean grain) typically accumulate below the cleaning unit near the bottom of the combine harvester. From there, a grain elevator transports the clean grain to a large grain tank that holds the harvested grain until it is unloaded into, for example, a trailer or truck that takes the harvested grain from the field. Straw, husks, and other non-grain materials typically fall or are scattered on the field. Non-grain materials are generally referred to as MOG (materials other than grain).

[0003] Ideally, a combine harvester will extract all the grains from the crop and load them into the grain tank without simultaneously picking up husks, straw, and other non-grain materials. In practice, an optimum must be found between maximizing the amount of grain harvested (per square meter of field and / or per minute) and minimizing the amount of broken grains or MOG in the grain tank. Many operating parameters of the combine harvester can be controlled to achieve this optimum. For example, such operating parameters include travel speed, threshing rotor speed, rotor cage clearance, cleaning screen openings, and cleaning fan speed. Many control settings that promote high grain yields can also result in grain breakage and large amounts of MOG. Conversely, control settings that minimize grain breakage and MOG generally result in lower grain yields. The optimum control settings are difficult to determine and constantly change with, for example, changes in the crop, weather, and field conditions.

[0004] To improve the control of all operating parameters of a combine harvester, many different sensors are used. A particularly useful sensing system for this purpose today is the so-called grain cam, such as that disclosed in the international patent application published as WO 2006 / 010761A1, which periodically captures images of the grain in a clean grain elevator or of a grain sample taken from the clean grain elevator. The captured images are analyzed using standard and more advanced image recognition techniques to distinguish grains, straw fragments, broken grains, and husks. When the MOG content of the observed grain sample exceeds a predetermined limit, the combine harvester settings and travel speed can be adjusted.

[0005] Although the grain cam disclosed in WO 2006 / 010761 A1 is useful for determining the relative amounts of MOG or broken grains in a grain sample, it cannot distinguish all relevant types of MOG. More specifically, the known grain cam cannot distinguish between empty husk particles and husk particles that still retain a grain (commonly referred to as white caps). In the images captured by the grain cam, (empty) husk particles and white caps cannot be distinguished. This is a problem because excessive husks and excessive white cap content in a grain sample require very different countermeasures. For example, excessive husks can be avoided by increasing the fan speed of the cleaning fan. For example, excessive white caps can be avoided by adjusting the rotor speed of the threshing rotor and / or by reducing the gap between the threshing rotor and the rotor cage.

[0006] An object of the present invention is to solve one or more disadvantages associated with the prior art. Summary of the Invention

[0007] According to one aspect of the present invention, there is provided an apparatus for analyzing a grain sample, comprising a light source, an image sensor, and a controller. The light source is configured to illuminate the grain sample. The image sensor is for capturing an image of the grain sample. The controller is coupled to the image sensor for receiving the image of the grain sample therefrom and is configured to analyze the image to detect at least one material other than grain in the grain sample. The light source used in the apparatus according to the present invention is configured to illuminate the grain sample with a local light spot having a size smaller than the width of an average wheat grain.

[0008] When the local light spot hits an empty husk particle, at least a portion of the light will pass through the husk shell and reflect on the inner surface of the empty shell. Some of this reflected light (possibly after multiple internal reflections) will pass through the husk shell again and then be captured by the image sensor. Thus, the entire (or almost entire) husk particle lights up in the captured image of the grain sample. When the local light spot hits a white cap, i.e., a husk particle containing a grain, the grain absorbs the light that initially passes through the husk shell. This absorption of light by the grain prevents the light from penetrating deep into the husk particle and reflecting on the inner surface of the husk shell. Thus, in the captured image of the grain sample, only the local direct reflection of the light spot at the outer shell of the husk particle will light up. Therefore, it can be determined from the width of the reflection whether the husk particle contains a grain. This enables the apparatus according to the present invention to distinguish between empty husk particles and white caps.

[0009] To allow the grain to absorb most of the incident light, it is important that the size (diameter, length, and / or width) of the light spot is smaller than the width of an average grain. A larger light spot will illuminate the entire outer surface of the husk shell. Then the direct reflection on its outer surface will cause the entire husk shell to appear in the captured image, making it impossible to distinguish between empty husk particles and white caps.

[0010] The projection of the light spot on the cereal sample is preferably circular or substantially circular, but light spots of different shapes can be used as an alternative. For a circular light spot, the light spot size is defined by the diameter of the light spot. For the present invention, the "size" of any non-circular light spot is defined herein as the maximum available dimension identifiable in the light spot. This maximum dimension will typically be the maximum of the length and width of the light spot.

[0011] Different cereal crops and cereal crop varieties may have grains of different sizes. Preferably, the light spot size is smaller than the average grain of the smallest available cereal crop variety, so that the device can be used for various crops. Since the device according to the present invention will be mainly used for wheat, the light spot size is preferably at least smaller than the average size of wheat grains. Such light spots can also be used to detect white caps in other cereal crop types and crop type varieties (such as rye, triticale, oats, soybeans or rice). Preferably, the diameter of the local light spot is less than 5 mm, more preferably less than 3 mm, 2 mm, 1 mm or 0.5 mm.

[0012] In a preferred embodiment, the light source includes a laser source for generating a local light spot. Alternatively, a highly focused LED or other type of light source can be used.

[0013] In order to be able to detect white caps at different positions in the cereal sample, the light source is preferably configured to illuminate the cereal sample with a plurality of local light spots, the sizes of which are smaller than the width of the average wheat grain. The plurality of light spots can be arranged, for example, along at least a substantially straight line and / or in a grid pattern.

[0014] Alternatively or additionally, the light source can be configured to move the local light spot relative to the cereal sample. This can be achieved by controlling the position of the light spot or by moving the cereal sample. For example, a single light spot or a horizontal line with a plurality of light spots can be aimed at a fixed position while the cereal sample moves past that position. The movement of the cereal sample can be caused at least in part by gravity or by the upward transport of the cereal in a grain elevator.

[0015] In addition to the local light spot, the light source can also be configured to illuminate the cereal sample with a wide light spot, the size of which is several times larger than the length of the average wheat grain. The wide light spot can have a diameter of at least 5 cm. Preferably, the wide light spot illuminates the entire cereal sample or at least most of the cereal sample. The image captured using such a wide light spot can be processed in a known manner and provide information about the position of the husks in the cereal sample. This information can then be used to direct the local light spot to the position of the husk where the husk is detected. By using the local light spot, it can then be determined whether the detected husk represents an empty husk particle or a white cap. Alternatively, the information obtained from the image captured using the wide light spot can be used to limit the image processing area to those areas where the detected husks are particles.

[0016] According to another aspect of the present invention, there is provided a combine harvester including an apparatus for analyzing a grain sample as described above. The combine harvester may include a clean grain elevator for conveying clean grain to a grain tank, and the apparatus for analyzing a grain sample is arranged to capture an image of the grain in the clean grain elevator or a bypass section thereof.

[0017] According to another aspect of the present invention, there is provided a method for analyzing a grain sample using the apparatus as described above. The method includes the following steps:

[0018] Illuminating the grain sample with a local light spot having a size smaller than the average wheat grain width using a light source,

[0019] Capturing at least one image of the grain sample using an image sensor,

[0020] Analyzing the at least one captured image to determine the reflection of the local light spot at the grain sample, and

[0021] Detecting materials other than the grain in the grain sample based on the determined reflection.

[0022] According to another aspect of the present invention, there is provided a computer program including instructions which, when executed by a computer, cause the computer to execute the method as described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Embodiments of the present invention will now be described by way of example with reference to the accompanying drawings.

[0024] Figure 1 A combine harvester is shown in which the grain cam according to the present invention can be advantageously used.

[0025] Figure 2 A clean grain elevator having a grain cam according to the present invention is schematically shown.

[0026] Figure 3 An image captured by Figure 2 the grain cam is shown, in which the grain sample is illuminated with a wide light spot.

[0027] Figure 4 A reflection curve of a chaff particle and a white cap when illuminated with a local light spot is shown.

[0028] Figure 5 An image captured by Figure 2 the grain cam is shown, in which the grain sample is illuminated with a local light spot.

[0029] Figure 6a A photograph of an empty chaff particle illuminated with a local light spot is shown.

[0030] Figure 6bA photograph of white cap particles illuminated by a local light spot is shown.

[0031] Figure 7 A flowchart showing an embodiment of the method according to the present invention is shown. Detailed Description

[0032] Figure 1 An agricultural harvester in the form of a combine harvester 10 is shown, which generally includes front and rear round engaging wheels 14, 16, a cutter bar 18, a feeder 20, an operator's cab 22, a threshing and separating system 24, a cleaning system 26, a grain tank 28, and a discharge pipe 30. It should be appreciated that although the agricultural harvester is shown as a combine harvester 10, the agricultural harvester according to the present invention can be implemented in any configuration that allows for the harvesting of crop material, such as a conventional combine harvester (which does not have a rotor), a rotary combine harvester, a hybrid combine harvester, a chopping harvester, etc.

[0033] The cutter bar 18 is mounted at the front of the combine harvester 10 and includes a cutter bar 34 for cutting crops from the field during forward movement of the combine harvester. A rotatable reel 36 feeds the crops into the cutter bar 18, and a double auger 38 laterally feeds the cut crops from each side into the feeder 20. The feeder 20 transports the cut crops to the threshing and separating system 24.

[0034] The threshing and separating system 24 is axial flow type and includes a threshing rotor 40 that is at least partially located within a threshing concave 42 and is rotatable. The threshing concave may take the form of a perforated concave. Under the action of the threshing rotor 40 within the threshing concave 42, the grains from the cut crops are threshed and separated from the MOG. The larger elements of the MOG, such as stems and leaves, are discharged from the rear of the combine harvester 10 without passing through the holes in the threshing concave 42. The grains and smaller elements of the MOG (hereinafter referred to as small MOG, such as chaff, dust, and straw) are small enough to pass through the perforations in the threshing concave 42 and then are discharged from the threshing and separating system 24.

[0035] The grains and small MOG that have successfully passed through the threshing and separating system 24 fall onto a preparation tray 44 and are conveyed to the cleaning system 26. The cleaning system includes a series of sieves and a cleaning fan 52. The series of sieves includes a pre-cleaning sieve 46, an upper sieve (or chaff sieve) 48, and a lower sieve (or shoe sieve) 50. The cleaning fan 52 generates an air flow that passes through the sieves 46, 48, 50, which impinges on the grains and small MOG on the sieves. The small MOG is generally lighter than the grains, so it separates from the grains when it is in the air. Subsequently, the small MOG is discharged from the combine harvester 10 via a straw hood 54.

[0036] The preparation pan 44 and the pre-cleaning sieve 46 swing in a front-to-back manner to convey the grain and small MOG to the upper surface of the upper sieve 48. The upper sieve 48 is vertically arranged above the lower sieve 50 and also swings in a front-to-back manner, so that the grain and small MOG are scattered on the two sieves 48, 50, while also allowing the cleaned grain to pass through the openings in the sieves 48, 50 under the action of gravity.

[0037] The cleaned grain falls into the cleaned grain auger 56, which is positioned below and in front of the lower sieve 50 and spans the width of the combine harvester 10. The cleaned grain auger 56 laterally conveys the cleaned grain to the vertical grain elevator 60, which is arranged to convey the cleaned grain to the grain tank 28. Once inside the grain tank 28, the grain tank auger 68 at the bottom of the grain tank laterally conveys the cleaned grain inside the grain tank 28 to the unloading pipe 30 for discharge from the combine harvester 10.

[0038] Figure 2 is schematically shown Figure 1 the cleaned grain elevator 60 of the combine harvester 10, which is equipped with a grain cam 100 according to an embodiment of the present invention. The grain elevator 60 includes a chain or belt 601, to which paddles 602 are attached. The chain or belt 601 is driven to move the paddles 602 upward on one side of the grain elevator 60 and downward on the other side. When moving upward, the paddles 602 carry the cleaned grain (not shown) obtained from the cleaned grain auger 56 below the lower sieve 50. At the top of the grain elevator 60, this cleaned grain is then transferred to a screw conveyor for introducing the grain into the grain tank 28, and then the paddles 602 for conveying the grain start their downward journey.

[0039] In the upward section of the grain elevator 60, there is a bypass 710. During operation, a portion of the grain conveyed upward by the grain elevator paddles 602 is directed or can be directed into the bypass 710. The bypass 710 includes a window 615. When the bypass 710 is filled with grain, the grain can be seen from the outside of the bypass 710 through the window 615. The grain cam 100 is attached to the bypass 710 in such a way that the image sensor or camera 120 of the grain cam 100 can capture an image of the grain sample visible behind the window 615. In addition to the camera 120, the grain cam 100 also includes one or more light sources 130, 140 for illuminating the grain sample when capturing an image of the grain sample behind the window 615. A controller 150 is provided for controlling the light sources 130, 140 and the camera 120, and for processing the image data obtained by the camera 120.

[0040] Communication line 155 couples the grain cam controller 150 to the central controller of the combine harvester 10 such that relevant control settings of the combine harvester 10 can be adjusted based on the results of the image analysis performed by the grain cam controller 150. Such adjustment of the combine harvester control settings can be fully automatic, user-initiated, or a combination thereof. Alternatively, the grain cam control function and / or the image data processing function implemented in the grain cam processor 150 is at least partially performed by the central controller of the combine harvester 10 or some other controller operably coupled to the grain cam 100.

[0041] According to the present invention, the light sources 130, 140 at least include a local spot source 140 configured to illuminate the grain sample with a local spot having a size smaller than the average wheat grain width. The local spot source 140 can include a laser. Alternatively, a highly focused LED or other type of light source can be used. In addition to the local spot source 140, a wide spot source 130 can be provided to illuminate the grain sample with a wide spot having a size several times larger than the average wheat grain length. The wide spot can have a diameter of at least 5 cm. Preferably, the wide spot illumination can cover all or at least most of the grain sample observable through the window 615. The wide spot source 130 can include, for example, a plurality of LEDs or one or more incandescent light sources. Images captured using such a wide spot can be processed in a known manner and information about the position of the chaff in the grain sample can be provided.

[0042] Note that although the grain cam 100 shown here is configured to capture images of grain samples in the bypass section 710 of the clean grain elevator 60 of the combine harvester 10, the grain cam 100 can equally be applicable to other locations where analysis of grain samples is required. For example, such other locations can include different locations in the combine harvester 10, a barn, or a laboratory environment.

[0043] Figure 3 An image captured by Figure 2 the grain cam 100 is shown, in which the grain sample is illuminated by the wide spot source 130. The images captured under these illumination conditions are the same or similar to the images that can be captured by a grain cam as described, for example, in WO 2006 / 010761A1. As Figure 3 visible therein, the image shows a plurality of grains 310 and the husks 320, 330 of four chaffs, which are visible through the window 615. In fact, other types of MOGs, such as straw particles, may also appear in the image. The image shows the light reflected from the outer surfaces of the grains 310 and the husks 320, 330 recorded by the camera 120. The white caps are the husks that still contain grains.

[0044] From the outside, the husks of empty husk grains (also referred to herein as "husk particles" 320) look very similar if not identical to the white caps 330. This is a problem because excessive husks and excessive white cap content in a grain sample require very different countermeasures. For example, excessive husks can be avoided by increasing the fan speed of the cleaning fan 52. For example, excessive white caps can be avoided by adjusting the rotor speed of the threshing rotor 40 and / or by reducing the gap between the threshing rotor 40 and the rotor cage 42.

[0045] According to the present invention, this problem is solved by adding a local light spot source 140 to the grain cam 100. When the local light spot hits an empty husk particle 320, at least a portion of the light passes through the husk and reflects off the inner surface of the empty husk 320. Some of this reflected light (possibly after multiple internal reflections) will pass through the husk of the husk 320 again and then be captured by the camera 120. Thus, the entire (or almost entire) husk particle 320 lights up in the captured image of the grain sample. When the local light spot hits a white cap 330 (i.e., a husk particle containing a grain), the grain absorbs or very locally reflects the light that initially passes through the husk. This absorption or very local reflection of the light by the grain prevents the light from penetrating deep into the husk particle and reflecting off the inner surface of the husk. Thus, in the captured image of the grain sample, only the very local direct reflection of the light spot at the outer shell of the husk particle lights up. Thus, it can be determined from the width of the reflection whether the husk particle contains a grain. This enables the grain cam 100 to distinguish between empty husk particles 320 and white caps 330.

[0046] To illustrate this difference in reflection behavior, Figure 4 shown are the reflection curves of an empty husk particle 320 and a white cap 330 when illuminated by a local light spot. As can be seen in this figure, both the empty husk particle 320 and the white cap (330) show a reflection peak (indicated in arbitrary units) caused by the direct reflection of the light spot on the outer surface of the husk. If the husk encloses a grain, as in the case of the white cap 330, then most of the remaining light is absorbed by the grain. If the husk is empty, the husk will glow on almost its entire outer surface.

[0047] To allow the grain to absorb most of the incident light, it is important that the size (diameter, length, and / or width) of the light spot is smaller than the width of the average grain. A larger light spot will illuminate the entire outer surface of the husk. Then the direct reflection on its outer surface will cause the entire husk to appear in the captured image, making it impossible to distinguish between empty husk particles 320 and white caps 330. Preferably, the diameter of the local light spot is less than 5 mm, more preferably less than 3 mm, 2 mm, 1 mm, or 0.5 mm.

[0048] Figure 5 shown is by Figure 2Images captured by the cereal 100 cam, where the cereal sample is illuminated by the local light spot 500. Here, it can be seen that two empty husk grains 320 glow on their entire outer surface, while the white caps 330 only show a small direct reflection of the local light spot 500.

[0049] In order to be able to detect the white caps 330 at different positions in the cereal sample and obtain an image such as Figure 5 as shown in, the local light spot source 140 is preferably configured to illuminate the cereal sample with a plurality of local light spots. The plurality of light spots can be arranged, for example, along at least substantially a straight line and / or in a grid pattern.

[0050] Alternatively or additionally, the local light spot source 140 can be configured to move the local light spot relative to the cereal sample. This can be achieved by controlling the position of the light spot or by moving the cereal sample. For example, a single light spot or a horizontal line with multiple light spots can be aimed at a fixed position while the cereal sample moves past that position. For example, this can be achieved by gradually or continuously lowering the platform carrying the cereal sample while the camera 120 captures a series of images of the cereal sample. Alternatively, the bypass 710 has a controllable outlet positioned at a location below the window 615. When the bypass 710 is filled with cereal, the outlet remains closed. Then a first image can be captured with the outlet still closed. Then subsequent images can be captured as the cereal exits the bypass 710 via the outlet, resulting in the cereal sample temporarily moving gradually downward along the bypass window 615. Other mechanical solutions for achieving a similar movement of the cereal sample along the bypass window 615 will be apparent to any person skilled enough in the art. Alternatively, when the camera is directly mounted on the cereal elevator 60, the cereal is lifted upward by the elevator, also generating a relative movement of the local light spot on the cereal surface during the image capture process.

[0051] Figure 6a and Figure 6b respectively show photos of empty husk grains and white cap grains in a larger cereal sample illuminated by the local light spot. These two photos confirm what has been described above with reference to Figure 5 the images. In both photos, the direct reflection of the local light spot can be seen. In addition, Figure 6a clearly shows how the entire empty husk grain glows due to internal reflection of the light inside the empty husk grain. In Figure 6b , it can be seen how most of the light is absorbed by the grain inside the white cap grain, thus significantly reducing internal reflection. Therefore, the white cap grains mainly glow near the area hit by the local light spot. As described above, Figure 6a the difference in this reflection pattern between the empty husk grains in the photo of Figure 6b and the white cap grains in the photo of is used by the device and method according to the present invention to detect the white cap grains in the cereal sample.

[0052] Figure 7 FIG. 3 shows a flowchart of an embodiment of a method according to the present invention. The method starts with a wide illumination step 710, in which a wide spot light source 130 is used to illuminate at least most, and preferably all, of the cereal sample (as long as it is visible through the window 615). In a first image capture step 720, the camera 120 then captures at least one image of the cereal sample. The image obtained in this first image capture step 720 may look like, for example, the image shown in FIG. Figure 3 FIG.

[0053] Using known image processing techniques such as filtering, segmentation, edge detection, and thresholding, the captured image is then analyzed in a first image analysis step 730. In this first image analysis step 730, the grains 310, the husks 320, 330 of the husk, the straw particles, and other types of MOG can be identified. In a preferred embodiment, a trained neural network and other artificial intelligence (AI) algorithms can be used to identify and count the different components of the cereal sample. A training data set for such AI algorithms can be obtained using previously analyzed images.

[0054] In a subsequent local illumination step 740, a local spot source 140 is used to illuminate the cereal sample with a local spot having a size smaller than the average wheat grain width. Preferably, a plurality of local spots are used to illuminate the cereal sample simultaneously at a corresponding plurality of different positions. Then, in a second image capture step 750, the camera 120 captures at least one image of the cereal sample. The image obtained in this second image capture step 750 may look like, for example, the image shown in FIG. Figure 5 FIG.

[0055] In a second image analysis step 760, the one or more images obtained in the second image capture step 750 are analyzed to detect the empty husk particles 320 and the white caps 330 in the cereal sample. Similarly for this second image analysis step 760, standard and more advanced image processing techniques can be used. If an AI algorithm is used for this second image analysis step 760, then a training data set can be obtained again using previously analyzed images. Such training data can be generated using a plurality of controlled cereal samples, in which the empty husk particles 320 and / or the white caps 330 are marked before capturing and analyzing the images of the cereal sample. Preferably, techniques are used such that the marking cannot be detected by the camera 120 of the cereal cam 100. For example, radioactive or UV reflective dyes can be used to mark the white caps 330.

[0056] In a preferred embodiment, the result of the first image analysis step 730 is used as the input to the second image analysis step 720. In the case where the result of the first image analysis step 730 can indicate the positions of the husks 320, 330 of the chaff found in the cereal sample, then the second image analysis step 760 can be used to distinguish between the empty husk particles 320 and the white caps 330. Note that the first two steps of the method shown in FIG. 6 are optional and thus not necessary for the detection of the white caps. The white caps 330 and the empty husk particles 320 can be detected with a cereal cam 100 that includes only the local light spot source 140 without the wide light spot source 130.

Claims

1. An apparatus (100) for analyzing a grain sample, the apparatus (100) comprising: Light sources (130, 140) for illuminating the grain sample, An image sensor (120) for capturing an image of the grain sample, and A controller (150) coupled to the image sensor (120) for receiving the image of the grain sample therefrom, wherein, The light source (140) is configured to illuminate the grain sample with a local light spot having a size smaller than the width of an average wheat grain (310), The controller (150) is configured to: Analyze the image to determine a reflection curve of the local light spot at the grain sample; and Detect at least one material (320, 330) other than the grain in the grain sample based on the reflection width of the determined reflection curve; and The at least one material other than the grain includes empty husk particles or white caps, and the white caps are husk particles still holding grains.

2. The apparatus (100) according to claim 1, wherein the local light spot has a diameter smaller than 5 mm.

3. The apparatus (100) according to claim 1 or 2, wherein the light source (140) includes a laser source.

4. The apparatus (100) according to claim 1, wherein the light source (140) is configured to illuminate the grain sample with a plurality of local light spots having a size smaller than the width of an average wheat grain (310).

5. The apparatus (100) according to claim 4, wherein the plurality of local light spots are arranged along at least a substantially straight line.

6. The apparatus (100) according to claim 4 or 5, wherein the plurality of local light spots are arranged in a grid pattern.

7. The apparatus (100) according to claim 1 or 2, wherein the light source (140) is configured to move the local light spot relative to the grain sample.

8. The apparatus (100) according to claim 1 or 2, wherein the light source (130) is further configured to illuminate the grain sample with a wide light spot having a size several times larger than the length of an average wheat grain.

9. The apparatus (100) according to claim 8, wherein the wide light spot has a diameter of at least 5 cm.

10. A combine harvester (10) comprising the apparatus (100) for analyzing a grain sample according to any of the preceding claims.

11. The combine harvester (10) according to claim 10, comprising a clean grain elevator (60) for conveying clean grain to a grain tank (28), and the apparatus (100) for analyzing a grain sample is arranged to capture an image of the grain in the clean grain elevator (60) or its bypass section (710).

12. A method for analyzing a grain sample using the apparatus (100) according to any one of claims 1 to 9, the method comprising the steps of: Illuminating the grain sample with a local light spot having a size smaller than the width of an average wheat grain using the light source (140), Capturing at least one image of the grain sample using the image sensor (120), Analyzing the at least one captured image to determine a reflection curve of the local light spot at the grain sample, and Detecting materials other than the grain in the grain sample based on the reflection width of the determined reflection curve.

13. The method according to claim 12, further comprising classifying the type of material other than the grain based on the determined reflection.

14. The method according to claim 13, further comprising illuminating the grain sample with a wide light spot using a light source (130) having a size several times the length of an average wheat grain.

15. The method according to claim 14, wherein a trained neural network and / or other artificial intelligence (AI) algorithms are used to detect materials other than the grain.

16. A computer program product comprising instructions that, when executed by a computer, cause the computer to perform the method according to any one of claims 12 to 15.

Images