Cereal multi-feature recognition system

Abstract

The utility model discloses a cereal multi-feature recognition system, which comprises a first photographing assembly and a second photographing assembly. The first photographing assembly is located on one side of the transparent rotating disc and comprises a plurality of first cameras, the first cameras are obliquely arranged at intervals along the circumference around the vertical line of the first station detection point, and the first cameras all align at the first station detection point. The distances from the plurality of first cameras to the first station detection point are respectively the optimal focal lengths of the corresponding first cameras; the second photographing assembly is located on the other side of the transparent rotary disc and comprises a second camera, the second camera is perpendicularly aligned with the second station detection point, and the distance between the second camera and the second station detection point is the optimal focal length of the second camera. According to the utility model, the appearance quality detection result of grain samples, especially large grain samples such as corn and soybeans, can be improved, and calculation resources can be saved.

Description

Technical Field

[0001] The utility model relates to the technical field of grain detection instruments, in particular to a grain multi-feature recognition system. Background Art

[0002] The appearance quality characteristics of grains often require observation of the appearance morphology data of grains. Under traditional conditions, it is usually dependent on quality inspectors to directly observe the appearance and judge the appearance quality of grains based on the main characteristics of the grains and the presence or absence of defects. Therefore, the accuracy of the test results is greatly affected by personal factors such as the experience and professional quality of the quality inspectors. With the rapid development of science and technology, visual recognition technology is becoming more and more widely used in agriculture. Use a visual recognition camera on a dedicated grain testing instrument to take pictures of grains, identify detailed features, and judge the quality of grains based on the features.

[0003] The recognition module in the existing detection equipment usually has the following problems: First, the feature recognition capability is insufficient. In current technology, a camera is often used to shoot in a fixed position, so only the features of a single plane can be recognized. Or a device that can collect front and back pictures of grains can be used to design a double camera up and down. For small grains such as wheat, sorghum, and rice, they are elliptical in shape. Collecting the front and back sides can fully cover all the surface features of the grains. The front and back sides can not only collect the features of the front and back sides, but also the features of the side. This is determined by the size and shape of the grains. However, for large grains such as corn and soybeans, especially corn, they are polygonal three-dimensional shapes. The shape of corn grains is similar to a flat rectangular shape. The features are distributed in multiple directions. It is difficult for a single camera or two cameras to fully collect and recognize all epidermal features. Second, multiple features need to be matched. In other solutions using multiple cameras, a single camera is arranged at multiple detection positions, and each camera shoots different postures. When the grain passes through multiple detection positions, multiple features at different angles are photographed in sequence. Therefore, multiple features of the same grain may be taken at different times, and it is necessary to find matches between multiple photos of the same grain before analysis can be performed, which takes up a lot of computing resources and brings difficulties to subsequent algorithm processing. Utility Model Content

[0004] The utility model aims to solve at least one of the technical problems existing in the prior art. To this end, one purpose of the utility model is to propose a grain multi-feature recognition system that can improve the appearance quality detection results of grain samples, especially large-grain grain samples such as corn and soybeans, and save computing resources.

[0005] The grain multi-feature recognition system according to the embodiment of the utility model comprises:

[0006] A first photographing component, the first photographing component is located on one side of the transparent turntable, the first photographing component includes a plurality of first cameras, the plurality of first cameras are arranged obliquely along a circumference around a perpendicular line of a first station detection point, and the plurality of first cameras are all aimed at the first station detection point, and the distances from the plurality of first cameras to the first station detection point are respectively the optimal focal lengths of the corresponding first cameras;

[0007] The second camera assembly is located on the other side of the transparent turntable, and the second camera assembly includes a second camera, and the second camera is vertically aligned with the second workstation detection point, and the distance from the second camera to the second workstation detection point is the best focal length of the second camera.

[0008] According to the grain multi-feature recognition system of the embodiment of the utility model, the working process is: first, the grain sample particles are placed on a transparent turntable, and the grain sample particles will rotate with the transparent turntable. When the grain sample particles move to the first station detection point, the first camera component takes a clear image of the remaining surfaces of the grain sample particles at the first station detection point except the other side surface (such as the bottom surface) from one side (such as the top) of the transparent turntable from multiple different angles; when the grain sample particles pass through the second station detection point, the second camera component takes a clear image of the other side surface (such as the bottom surface) of the grain sample particles from the other side (such as the bottom) of the transparent turntable. The images taken by the first camera component at the first station detection point and the second camera component at the second station detection point can obtain all the epidermal features of the grain sample particles through post-processing by an algorithm, and then the appearance quality of the grain sample particles is judged according to all the epidermal features of the grain sample particles, which greatly improves the appearance quality detection results of grain samples, especially large-grain grain samples such as corn and soybeans.

[0009] The grain multi-feature recognition system of the embodiment of the utility model can take all-round photos of various grain particles, including large-grain grain particles such as corn and soybeans, greatly improving the appearance quality detection results of grain samples, especially large-grain grain samples such as corn and soybeans, and can reduce the complex process of identifying all the surface features of grain sample particles such as post-image matching, thereby saving computing resources.

[0010] In some embodiments, the first workstation detection point and the second workstation detection point coincide with each other as the same workstation detection point.

[0011] In some embodiments, the first camera assembly is located on the upper side of the transparent turntable, and the second camera assembly is located on the lower side of the transparent turntable.

[0012] In some embodiments, the number of the plurality of first cameras is 2 to 4.

[0013] In some embodiments, the tilt angles of the first cameras are all between 15° and 25°.

[0014] In some embodiments, the plurality of first cameras are located at the same height, and the plurality of first cameras are at the same distance from the first workstation detection point.

[0015] In some embodiments, the plurality of first cameras are distributed at equal intervals along a circumferential direction.

[0016] In some embodiments, the first photographing component further includes a first fill light plate, and the first fill light plate is disposed between the plurality of first cameras and the transparent turntable.

[0017] In some embodiments, the second camera assembly further includes a second fill light plate, and the second fill light plate is disposed between the second camera and the transparent turntable.

[0018] In some embodiments, a bracket is further included, and the first camera component and the second camera component are both mounted on the bracket.

[0019] Additional aspects and advantages of the present invention will be given in part in the following description, and in part will become apparent from the following description, or will be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0021] Figure 1 It is a schematic diagram of the overall structure of the grain multi-feature recognition system of the embodiment of the utility model.

[0022] Figure 2 This is a schematic diagram of the installation of the first camera of the grain multi-feature recognition system of the utility model embodiment.

[0023] Reference numerals: grain multi-feature recognition system 1000 ; first photographing component 100 ; first camera 101 ; first fill light plate 102 ; second photographing component 200 ; second camera 201 ; second fill light plate 202 ; transparent turntable 300 ; bracket 400 . DETAILED DESCRIPTION

[0024] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0025] Combine the following Figure 1and Figure 2 To describe the grain multi-feature recognition system 1000 according to an embodiment of the present utility model.

[0026] like Figure 1 and Figure 2 As shown, the grain multi-feature recognition system 1000 according to the embodiment of the utility model includes a first camera component 100 and a second camera component 200. The first camera component 100 is located on one side of the transparent turntable 300, and the first camera component 100 includes a plurality of first cameras 101, which are arranged obliquely along the circumference around the vertical line of the first station detection point, and the plurality of first cameras 101 are all aimed at the first station detection point, and the distances from the plurality of first cameras 101 to the first station detection point are the best focal lengths of the corresponding first cameras 101; the second camera component 200 is located on the other side of the transparent turntable 300, and the second camera component 200 includes a second camera 201, which is vertically aimed at the second station detection point, and the distance from the second camera 201 to the second station detection point is the best focal length of the second camera 201.

[0027] Specifically, the first camera assembly 100 is located on one side of the transparent turntable 300, wherein the transparent turntable 300 can rotate to facilitate the transportation of the grain sample particles to the first station detection point and the second station detection point. The first camera assembly 100 module includes a plurality of first cameras 101, and the plurality of first cameras 101 are arranged obliquely along the circumference around the vertical line of the first station detection point, and the plurality of first cameras 101 are all aimed at the first station detection point, and the distances from the plurality of first cameras 101 to the first station detection point are the best focal lengths of the corresponding first cameras 101. In other words, the first camera assembly 100 can clearly photograph the remaining surfaces of the grain sample particles at the first station detection point except the other side surface (such as the bottom surface) from one side (such as the top) of the transparent turntable 300 from multiple different angles, and can capture the appearance features of the remaining surfaces of the grain sample particles except the other side surface (such as the bottom surface), which is conducive to improving the accuracy of the grain particle appearance detection results. The multiple first cameras 101 identify the same grain sample particles at the first workstation detection point, that is, photograph the grain sample particles at the same time, which can reduce the complex process of subsequent image matching and save computing resources.

[0028] The second camera assembly 200 is located on the other side of the transparent turntable 300. The second camera assembly 200 includes a second camera 201. The second camera 201 is vertically aligned with the second station detection point. The distance from the second camera 201 to the second station detection point is the best focal length of the second camera 201. In other words, the second camera 201 can clearly capture the image of the other side surface (such as the bottom surface) of the grain sample particles that arrive at the second station detection point, and can capture the appearance characteristics of the other side surface (such as the bottom surface) of the grain sample particles. Therefore, the images captured by the second camera 201 and the images captured by the multiple first cameras 101 can be processed by the post-processing algorithm to obtain all the epidermal characteristics of the grain sample particles, and then judge the appearance quality of the grain sample particles based on all the epidermal characteristics of the grain sample particles, which greatly improves the appearance quality test results of grain samples, especially large grain samples such as corn and soybeans.

[0029] like Figure 1 and Figure 2 As shown, according to the grain multi-feature recognition system 1000 of the embodiment of the utility model, the working process is as follows: first, the grain sample particles are placed on the transparent turntable 300, and the grain sample particles will rotate with the transparent turntable 300. When the grain sample particles move to the first station detection point, the first camera component 100 takes a clear image of the remaining surfaces of the grain sample particles at the first station detection point except the other side surface (such as the bottom surface) from one side (such as the top) of the transparent turntable 300 from multiple different angles; when the grain sample particles pass through the second station detection point, the second camera component 200 takes a clear image of the other side surface (such as the bottom surface) of the grain sample particles from the other side (such as the bottom) of the transparent turntable 300. The images taken by the first camera component 100 at the first station detection point and the second camera component 200 at the second station detection point can obtain all the epidermal features of the grain sample particles through post-processing by the algorithm, and then judge the appearance quality of the grain sample particles according to all the epidermal features of the grain sample particles, which greatly improves the appearance quality detection results of grain samples, especially large-grain grain samples such as corn and soybeans.

[0030] The grain multi-feature recognition system 1000 of the embodiment of the utility model can take all-round photos of various grain particles, including large-grain grain particles such as corn and soybeans, greatly improving the appearance quality detection results of grain samples, especially large-grain grain samples such as corn and soybeans, and can reduce the complex process of identifying all the surface features of grain sample particles such as post-image matching, thereby saving computing resources.

[0031] In some embodiments, the first station detection point and the second station detection point overlap as the same station detection point. In this way, the first camera component 100 and the second camera component 200 can simultaneously capture images of grain sample particles, and the grain particle detection effect is better; and the first camera component 100 and the second camera component 200 are arranged on the upper and lower sides of the same station detection point, so that the device is small in size and can be easily processed by the post-processing algorithm. It should be noted that if the first station detection point and the second station detection point do not overlap, it is sufficient to ensure that images of all sides of the grain particles can be captured.

[0032] In some embodiments, the first camera assembly 100 is located on the upper side of the transparent turntable 300, and the second camera assembly 200 is located on the lower side of the transparent turntable 300. In this way, the first camera assembly 100 can clearly photograph the surfaces other than the bottom surface of the grain sample particles at the first station detection point from above the transparent turntable 300 from multiple different angles, and can capture the appearance characteristics of all surfaces other than the bottom surface of the grain sample particles, which is conducive to improving the accuracy of the grain particle appearance detection results; and multiple first cameras 101 identify the same grain sample particles at the first station detection point, that is, photograph the grain sample particles at the same time, which can reduce the complex process of later image matching and save computing resources. The second camera 201 can clearly capture the bottom surface image of the grain sample particles arriving at the second station detection point, and can capture the appearance characteristics of the bottom surface of the grain sample particles. Thus, the images captured by the second camera 201 and the images captured by the multiple first cameras 101 can be processed by the post-processing algorithm to obtain all the epidermal features of the grain sample particles, and then the appearance quality of the grain sample particles can be judged based on all the epidermal features of the grain sample particles, which greatly improves the appearance quality detection results of grain samples, especially large grain samples such as corn and soybeans. It should be explained that the second camera component 200 can be set on the upper side of the transparent turntable 300, and the first camera component 100 can be set on the lower side of the transparent turntable 300, and the same shooting effect can also be obtained.

[0033] In some embodiments, the number of the plurality of first cameras 101 is 2 to 4, that is, the first cameras 101 may have 2 (eg, Figure 2 ), 3 or 4 (as shown Figure 1 As shown), when detecting large-grain cereal particles, the appearance detection results of the large-grain cereal particles can be greatly improved. It is suitable for the appearance quality detection of large-grain cereal particles such as corn and soybeans, and is also suitable for the appearance quality detection of small-grain cereal particles such as wheat, sorghum, and rice.

[0034] In particular, when the grain sample particles are corn, since corn is a flat rectangular three-dimensional structure, in order to fully capture the corn epidermis features, it is necessary to capture the features of the six sides of the corn. Ideally, a picture of each of the six sides of the corn needs to be captured for algorithm recognition. However, after many experiments, it was found that by setting five cameras in the grain multi-feature recognition system 1000, that is, four first cameras 101 and one second camera 201, all epidermis features of the corn grains can be captured.

[0035] When the grain sample particles are soybeans, since soybeans are large particles and have a nearly circular three-dimensional shape, setting three first cameras 101 or setting four first cameras 101 can capture all the surface characteristics of the soybeans.

[0036] When the grain sample particles are wheat, sorghum, and rice, since wheat, sorghum, and rice all present small oval three-dimensional features, setting two first cameras 101 can collect all the surface features of the grain sample particles.

[0037] In summary, the grain multi-feature recognition system 1000 of the embodiment of the utility model is suitable for the appearance quality detection of various grain particles, such as corn, soybeans, wheat, rice, sorghum and other grain particles in different forms, and can be easily integrated into different grain detection equipment and automatic grain detection lines.

[0038] In some embodiments, the tilt angles of the multiple first cameras 101 are all 15° to 25°. It is understandable that square particles and large particles require an increased tilt angle to see more areas, and small particles require a reduced tilt angle for higher clarity. Specifically, the defect features of small-grain cereals are relatively small and require clearer shooting, so the focus center is focused on the small particles, such as sorghum, which requires a small tilt angle; large-grain cereals have large defect features and do not need to consider the out-of-focus problem, but require a larger viewing angle, because a smaller tilt angle makes it impossible to see the defective area on the side, such as corn, soybeans, etc., which require a large angle; similarly, medium-grain cereals are between the two and require a moderate angle, such as wheat.

[0039] For example, Figure 2 As shown, after actual testing, when the grain sample particles are corn, when the angle α between the first camera 101 and the vertical angle is 20°, the first camera 101 can better cover the epidermal features of the upper surface of the corn and the epidermal features of the surrounding sides, so as to collect all the epidermal features of the corn particles.

[0040] In some embodiments, multiple first cameras 101 are located at the same height, and the distances from multiple first cameras 101 to the first station detection point are the same, which is convenient for installation and focusing, and for later stitching of photos, reducing complex processes such as later image matching.

[0041] In some embodiments, the plurality of first cameras 101 are evenly spaced along the circumferential direction, and the image coverage angle is adaptively increased according to the shape characteristics of the grain sample particles, so that more features of the grain sample particles can be captured.

[0042] In some embodiments, the first camera assembly 100 further includes a first fill light plate 102, which is disposed between the first cameras 101 and the transparent turntable 300. The first fill light plate 102 is used to provide light source compensation when the first cameras 101 take photos, thereby enhancing the shooting effect of the first cameras 101.

[0043] In some embodiments, the second camera assembly 200 further includes a second fill light plate 202, which is disposed between the second camera 201 and the transparent turntable 300. The second fill light plate 202 is used to provide light source compensation when the second camera 201 takes pictures, which can enhance the shooting effect of the second camera 201.

[0044] In some embodiments, a bracket 400 is also included, and the first camera component 100 and the second camera component 200 are both mounted on the bracket 400, which can fix the first camera component 100 and the second camera component 200, so that the first camera component 100 and the second camera component 200 can photograph the grain sample particles from a suitable angle and distance.

[0045] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the utility model. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0046] Although the embodiments of the present invention have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A grain multi-feature recognition system, characterized in that: include: A first photographing component, the first photographing component is located on one side of the transparent turntable, the first photographing component includes a plurality of first cameras, the plurality of first cameras are arranged obliquely along a circumference around a perpendicular line of a first station detection point, and the plurality of first cameras are all aimed at the first station detection point, and the distances from the plurality of first cameras to the first station detection point are respectively the optimal focal lengths of the corresponding first cameras; The second camera assembly is located on the other side of the transparent turntable, and the second camera assembly includes a second camera, and the second camera is vertically aligned with the second workstation detection point, and the distance from the second camera to the second workstation detection point is the best focal length of the second camera.

2. The grain multi-feature recognition system according to claim 1, characterized in that: The first workstation detection point and the second workstation detection point coincide with each other as the same workstation detection point.

3. The grain multi-feature recognition system according to claim 1, characterized in that: The first photographing component is located on the upper side of the transparent turntable, and the second photographing component is located on the lower side of the transparent turntable.

4. The grain multi-feature recognition system according to claim 1, characterized in that: The number of the plurality of first cameras is 2 to 4.

5. The grain multi-feature recognition system according to claim 1, characterized in that: The inclination angles of the first cameras are all between 15° and 25°.

6. The grain multi-feature recognition system according to claim 1, characterized in that: The plurality of first cameras are located at the same height, and the distances from the plurality of first cameras to the first workstation detection point are the same.

7. The grain multi-feature recognition system according to claim 1, characterized in that: The plurality of first cameras are distributed at equal intervals along the circumferential direction.

8. The grain multi-feature recognition system according to any one of claims 1 to 7, characterized in that: The first photographing assembly further includes a first fill light plate, and the first fill light plate is arranged between the first cameras and the transparent turntable.

9. The grain multi-feature recognition system according to any one of claims 1 to 7, characterized in that: The second photographing assembly further includes a second fill light plate, and the second fill light plate is arranged between the second camera and the transparent turntable.

10. The grain multi-feature recognition system according to any one of claims 1 to 7, characterized in that: It also includes a bracket, and the first camera component and the second camera component are both installed on the bracket.

Images