CN115165772B - Device and method for detecting impurity content of grains based on hyperspectrum - Google Patents

Abstract

The invention provides a device and a method for detecting impurity content of grains based on hyperspectrum, wherein the device comprises a hyperspectral camera, a light source processing system, a light source and the like; the hyperspectral camera and the light source are positioned near the conveying device, and the conveying device is used for conveying seeds; the conveying device is provided with a detection area, the light source is used for generating incident light, and the incident light is aligned to the detection area; the hyperspectral camera is used for collecting reflected light reflected by the seeds; the light source processing system determines the reflected light intensity according to the collected reflected light of the seeds, the light source processing system determines the reflectivity according to the reflected light intensity and the incident light intensity, and the light source processing system determines the impurity content of the seeds in the detection period according to the determined reflectivity and the reflectivity of the pure seeds. The invention can realize the real-time sampling and monitoring of the impurity content of the seeds during crop harvesting, calculates the impurity content according to the principle that the light source absorption rate and the reflectivity of the seeds and the impurities are different, and improves the monitoring precision.

Description

A device and method for detecting impurity content of seeds based on hyperspectral

Technical Field

The invention relates to the field of agricultural machinery harvesting, and in particular to a device and a method for detecting impurity content of grains based on a hyperspectral spectrum.

Background Art

The detection of impurity content in crop grains of combine harvesters helps drivers understand the impurity content of crop grains in real time, and helps farmers judge the quality of harvest by the size of the impurity content, thereby adjusting the operating speed and working parameters of the cleaning device. Excessively high impurity content leads to an increase in the workload of subsequent grain processing and increases the processing cost of grain.

In the existing crop seed impurity detection devices and methods, most of them use industrial cameras in combination with image recognition to detect the impurity content of seeds. This method cannot exclude empty shells, resulting in measurement errors. The image needs to be segmented during measurement, which is a lot of work. When the device is sampling, it is mostly a single sampling. After the sample is released, the next sampling test is performed. This will inevitably cause incomplete sample release and accumulation, resulting in large errors in the test results.

Summary of the invention

In view of the deficiencies in the prior art, the present invention provides a device and method for detecting the impurity content of grains based on a hyperspectral spectrum, which can realize real-time sampling and monitoring of the impurity content of grains during crop harvesting, and calculate the impurity content according to the different absorption rates and reflectivity of the light source between the grains and the impurities, thereby improving the detection accuracy, helping the driver to understand the impurity content of the grains in real time, and helping the operator to judge the harvest quality by the size of the impurity content, thereby adjusting the operating speed and the working parameters of the cleaning device to improve the harvest quality.

The present invention achieves the above technical objectives through the following technical means.

A device for detecting impurity content of grains based on hyperspectral, comprising a hyperspectral camera, a light source processing system and a light source;

The hyperspectral camera and the light source are located near a conveying device, and the conveying device is used to convey grains. A detection area is provided on the conveying device, and the light source is used to generate incident light, and the incident light is aimed at the detection area. The hyperspectral camera is used to collect reflected light reflected by the grains. The light source processing system determines the intensity of the reflected light according to the collected reflected light of the grains, and the light source processing system determines the reflectivity according to the reflected light intensity and the incident light intensity. The light source processing system determines the impurity content of the grains within the detection period according to the determined reflectivity and the reflectivity of the pure grains.

Furthermore, at least one sampler is installed on the conveying device, and the sampler is conveyed synchronously with the conveying device; when the sampler enters the detection area, the light source processing system controls the hyperspectral camera to collect reflected light reflected by the grains in the sampler.

Furthermore, a partition is provided above the conveying device, dividing the detection device into a material receiving area and a monitoring area; the input end of the conveying device is located in the material receiving area, the output end of the conveying device is located in the monitoring area, and the hyperspectral camera, light source processing system and light source are located in the monitoring area; a gap is provided between the partition and the conveyed grains on the conveying device, so as to distribute the grains entering the monitoring area in sequence.

Furthermore, a light-absorbing coating is provided on the wall surface located in the monitoring area; and the partition is made of non-metallic material.

Furthermore, the incident light intensity generated by the light source is 1 _f , and the light source processing system determines the reflected light intensity as l _z according to the collected reflected light of the grains, and the reflectivity α is l _z /1 _f .

A detection method based on a hyperspectral seed impurity detection device comprises the following steps:

Determine the reflectance interval (α _min , α _max ) of pure kernels;

The reflected light reflected by the i-th sampled grain is collected by the hyperspectral camera, and the reflected light intensity l _zi of the i-th sample and the incident light intensity l _fi of the i-th sample are determined by the light source processing system;

The light source processing system obtains the reflectivity α _i of the i-th sampling according to the reflected light intensity l _zi of the i-th sampling and the incident light intensity l _fi of the i-th sampling; N is the total number of samplings, and the sampling period is T, i∈[1,N];

When α _i ≥ 0.9, the reflected light of the i-th sampling is not absorbed by the grains or impurities, and the invalid sampling times d = d + 1;

When α _i ＜0.9, if If the interval is α _i ∈ (α _min , α _max ), the number of times it contains impurities is j=j+1, and the number of samplings is i=i+1; if the interval is α i ∈ (α min , α max ), the number of samplings is i=i+1;

When i>N, the light source processing system calculates the impurity rate m=j/(N-d)×100%.

Furthermore, the reflectance interval (α _min , α _max ) of the pure grains is determined as follows:

The spectral wavelength of pure grains is collected by the hyperspectral camera to obtain the wavelength range of pure grain reflected light (λ _min , λ _max ). Assuming that the wavelength of incident light is λ ₀ , the initial reflectivity range of pure grains (α' _min , α' _max ) can be characterized as:

The initial reflectivity interval is corrected by the neural network:

The initial reflectivity interval is divided into four intervals: and

The neural network predicts that the distribution probability of pure grain light source reflectance falling into four intervals is A, B, C and D respectively;

Use A and B to correct α' _min , and the corrected α _min is:

Use C and D to correct α' _max , and the corrected α _max is:

Furthermore, the impurity rate is transmitted to the display through the RS485 communication system in the light source processing system for display and storage.

The beneficial effects of the present invention are:

1. The device and method for detecting impurity content in grains based on hyperspectral light described in the present invention can realize real-time sampling and monitoring of impurity content in grains when crops are harvested, and calculate the impurity content based on the different light source absorption and reflectivity of grains and impurities, thereby improving the monitoring accuracy.

2. The device and method for detecting impurity content of grains based on hyperspectral conditions described in the present invention divide the detection device into a receiving area and a monitoring area by a partition. On the one hand, this can prevent scattered light from other light sources in the receiving area from affecting the reflected light collected by the hyperspectral camera. On the other hand, a gap is provided between the bottom of the partition and the conveyed grains on the conveying device, so that the partition can be used as a scraper to convert disordered or accumulated grains on the conveying device into a set of grains arranged in a certain thickness, so that the grains entering the monitoring area are distributed in order, thereby achieving more accurate detection.

3. The device and method for detecting impurity content of grains based on hyperspectral data of the present invention can obtain a relatively accurate reflectance interval (α _min , α _max ) of pure grains by correcting the initial reflectance interval through a neural network.

4. The device and method for detecting impurities in grains based on hyperspectral light described in the present invention determine whether there are impurities in the grains by reflectivity. The present invention is applicable to the detection of impurities in grains of different crops, and does not require the measurement of moisture content, because pure grains with different moisture contents will have a determined reflectivity range before detection, and the reflectivity range will be corrected according to the reflectivity distribution probability, and empty shells can be automatically classified as impurities during measurement, further improving detection efficiency and accuracy.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. The drawings in the description are some embodiments of the present invention. For ordinary technicians in this field, it is obvious that other drawings can be obtained based on these drawings without paying any creative work.

FIG1 is a schematic diagram of a device for detecting impurity content in grains based on hyperspectral information according to the present invention.

FIG. 2 is a control principle diagram of the device for detecting impurity content in grains based on hyperspectral information according to the present invention.

FIG3 is a flow chart of the detection method of the device for detecting impurity content in grains based on hyperspectral information according to the present invention.

In the figure:

101- unloading grain stirring; 102- pulley; 103- sampler; 104- conveyor belt; 105- hyperspectral camera; 106- light source processing system; 107- halogen lamp.

DETAILED DESCRIPTION

The present invention is further described below in conjunction with the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

Embodiments of the present invention are described in detail below, examples of which 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 intended to be used to explain the present invention, and should not be construed as limiting the present invention.

In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "axial", "radial", "vertical", "horizontal", "inner", "outer" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

As shown in FIG1 , the hyperspectral-based impurity detection device of the present invention comprises a hyperspectral camera 105, a light source processing system 106, a halogen lamp 107 and a display; the hyperspectral camera 105 and the light source are located near the conveying device, one end of the conveying device is connected to the unloading grain stirring 101, the conveying device comprises a pulley 102 and a conveyor belt 104, and the conveyor belt 104 conveys the grain from the outlet of the unloading grain stirring 101 to the grain receiving bag. The conveying device is provided with a detection area, the halogen lamp 107 is used to generate incident light, and the incident light is aimed at the detection area; the hyperspectral camera 105 is used to collect reflected light reflected by the grain; the light source processing system 106 determines the reflected light intensity according to the reflected light of the collected grain, the light source processing system 106 determines the reflectivity according to the reflected light intensity and the incident light intensity, and the light source processing system 106 determines the impurity content of the grain within the detection period according to the determined reflectivity and the reflectivity of the pure grain. The display is used to display the impurity content of the grain within the period.

At least one sampler 103 is installed on the conveying device, and the sampler 103 is installed on the conveyor belt 104 and is conveyed synchronously with the conveying device; when the sampler 103 enters the detection area, the light source processing system 106 controls the hyperspectral camera 105 to collect the reflected light reflected by the grains in the sampler 103. In the embodiment, four samplers 103 are installed on the conveyor belt 104, and the samplers 103 move to the detection area in sequence, and the hyperspectral camera 105 collects the reflected light reflected by the grains on the sampler 103. After the sampler 103 rotates to the bottom with the pulley 102, the grains in the sampler 103 will automatically fall into the grain receiving bag, avoiding the waste of grains. A sensor can also be installed on the conveying device to identify whether the sampler 103 has entered the detection area, or directly use image recognition technology to determine whether the sampler 103 has entered the detection area.

The device for detecting impurity content of grains based on hyperspectral light of the present invention does not need to install the sampler 103 on the conveyor belt 104 , and only needs to set the interval time to detect the grains on the conveyor belt 104 that enter the detection area in each cycle.

A partition is provided above the conveying device, dividing the detection device into a receiving area and a monitoring area; the input end of the conveying device is located in the receiving area, the output end of the conveying device is located in the monitoring area, and the hyperspectral camera 105, the light source processing system 106 and the light source are located in the monitoring area; a light-absorbing coating is provided on the wall surface in the monitoring area; the purpose of providing the partition is to prevent the scattered light generated by other light sources in the receiving area from affecting the reflected light collected by the hyperspectral camera 105, which results in inaccurate reflected light intensity collected by the hyperspectral camera 105. Another more important purpose of providing the partition is to solve the problem that the grains may be piled up on the conveyor belt 104 or on the sampler 103 because the grain unloading and stirring 101 conveys the grains to the conveyor belt 104, which may cause the debris to be hidden under the piled grains and cannot be identified; and after the piled grains reach the monitoring area, the serious height difference on the surface makes it impossible for all the reflected light to enter the collection area of the hyperspectral camera 105. Therefore, a gap is provided between the bottom of the partition and the conveyed grains on the conveyor belt 104. The gap can generally be the thickness of 2-4 layers of grains, so that the accumulated grains become a grain layer of a certain height after passing through the partition, that is, it can be considered that the grains entering the monitoring area are distributed in order. To prevent the bottom of the partition from causing damage to the surface of the grains, resulting in an increase in the breakage rate, the partition is made of non-metallic material.

After the incident light is absorbed by the grains or impurities, the reflected light intensity will be different, the spectral information of each pixel in the reflection spectrum is extracted, and the hyperspectral feature extraction is performed through data dimension reduction and feature extraction. Based on the differences in the absorption and reflection of the spectrum by different crops, the light intensity after being absorbed and reflected by the crop grains or impurities is obtained. The incident light intensity generated by the light source is l _f , and the light source processing system 106 determines the reflected light intensity as l _z according to the collected reflected light of the grains, and the reflectivity α is l _z /l _f .

As shown in FIG2 , in the device for detecting impurity content of grains based on hyperspectral conditions of the present invention, the conveying device conveys the grains to the detection area, the hyperspectral camera 105 collects the reflected spectrum information of the grains, the light source processing system 106 converts the reflectivity and impurity content, and finally transmits the detected impurity content and other information to the display through the RS485 communication system.

As shown in FIG3 , the detection method of the seed impurity rate detection device based on the hyperspectral spectrum of the present invention comprises the following steps:

Determine the reflectance interval (α _min , α _max ) of pure grains, specifically:

The spectral wavelength of pure grains is collected by the hyperspectral camera 105 to obtain the wavelength range of pure grain reflected light (λ _min , λ _max ). Assuming that the wavelength of incident light is λ ₀ , the initial reflectivity range of pure grains (α' _min , α' _max ) can be characterized as:

The initial reflectivity interval is corrected by the neural network:

The initial reflectivity interval is divided into four intervals: and

The neural network predicts that the distribution probability of pure grain light source reflectance falling into four intervals is A, B, C and D respectively;

Use A and B to correct α' _min , and the corrected α _min is:

Use C and D to correct α' _max , and the corrected α _max is:

The reflected light reflected by the i-th sampled grain is collected by the hyperspectral camera 105, and the reflected light intensity l _zi of the i-th sample and the incident light intensity l _fi of the i-th sample are determined by the light source processing system 106;

The light source processing system 106 obtains the reflectivity α _i of the i-th sampling according to the reflected light intensity l _zi of the i-th sampling and the incident light intensity l _fi of the i-th sampling; N is the total number of samplings, and the sampling period is T, i∈[1,N];

When α _i ≥ 0.9, the reflected light of the i-th sampling is not absorbed by the grains or impurities, and the invalid sampling times d = d + 1;

When α _i ＜0.9, if If the interval is α _i ∈ α _min , α _max , the sampling number i = i + 1;

When i>N, the light source processing system 106 calculates the impurity rate m=j/(N-d)×100%.

After the sampling test is completed, it is determined whether to perform the next sampling test. If the sampling continues, the pulley continues to move. If the sampling is completed, the pulley stops moving.

The invention uses a hyperspectral camera to monitor the impurity content of crop seeds, which improves the monitoring accuracy and can sample and monitor crop seeds in real time. The seeds are transported to the hyperspectral system through a pulley for monitoring, and then the monitored seeds are automatically dropped into a sample receiving bag to avoid loss, thereby improving the monitoring efficiency of impurities. Whether there are impurities is determined by converting the intensity of the reflected light source into reflectivity.

It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

The series of detailed descriptions listed above are only specific descriptions of feasible embodiments of the present invention. They are not intended to limit the scope of protection of the present invention. All equivalent embodiments or changes that do not deviate from the technical spirit of the present invention should be included in the scope of protection of the present invention.

Claims (5)

1. The detection method based on the hyperspectral grain impurity rate detection device comprises a hyperspectral camera (105), a light source processing system (106) and a light source; the hyperspectral camera (105) and the light source are positioned near a conveying device, and the conveying device is used for conveying seeds; the conveying device is provided with a detection area, the light source is used for generating incident light, and the incident light is aligned to the detection area; the hyperspectral camera (105) is used for collecting reflected light reflected by the seeds; the light source processing system (106) determines the reflected light intensity according to the collected reflected light of the seeds, the light source processing system (106) determines the reflectivity according to the reflected light intensity and the incident light intensity, and the light source processing system (106) determines the impurity content of the seeds in the detection period according to the determined reflectivity and the reflectivity of the pure seeds; the method is characterized by comprising the following steps of:

The reflectance interval (α _min,α_max) of the pure grain was determined, specifically:

Collecting spectral wavelengths of pure kernels by the hyperspectral camera (105) to obtain a pure kernel reflected light wavelength range (lambda _min,λ_max), assuming that the wavelength of the incident light is lambda ₀, the initial reflectance interval (alpha' _min,α'_max) of the pure kernels can be characterized as

Correcting the initial reflectivity interval through a neural network:

The initial reflectance interval is divided into four intervals:

Predicting distribution probabilities of pure grain light source reflectivity falling into four intervals as A, B, C and D respectively through a neural network;

alpha' _min was modified with A and B to obtain modified alpha _min as:

alpha' _max was modified with C and D to obtain modified alpha _max as:

Collecting reflected light reflected by the ith sampled grain through the hyperspectral camera (105), and determining the ith sampled reflected light intensity l _zi and the ith sampled incident light intensity l _fi through a light source processing system (106);

The light source processing system (106) obtains the reflectivity alpha _i of the ith sample according to the reflected light intensity l _zi of the ith sample and the incident light intensity l _fi of the ith sample; n is the total sampling times, the sampling period is T, i epsilon [1, N ];

When alpha _i is more than or equal to 0.9, the reflected light of the ith sampling is not absorbed by seeds or impurities, and the sampling times d=d+1 are not valid;

When alpha _i is less than 0.9, if In the interval, the impurity times j=j+1, and the sampling times i=i+1; if α _i∈(α_min,α_max) interval, the sampling number i=i+1;

When i > N, the light source processing system (106) calculates the impurity content m=j/(N-d) ×100%.

2. The detection method based on the hyperspectral lower grain impurity rate detection device according to claim 1, wherein at least one sampler (103) is installed on the conveying device, and the sampler (103) synchronously conveys along with the conveying device; when the sampler (103) enters the detection zone, the light source processing system (106) controls the hyperspectral camera (105) to collect reflected light reflected by seeds in the sampler (103).

3. The detection method based on the hyperspectral kernel impurity rate detection device according to claim 1, wherein a partition plate is arranged above the conveying device, and the detection device is divided into a receiving area and a detection area; the input end of the conveying device is positioned in the receiving area, the output end of the conveying device is positioned in the detection area, and the hyperspectral camera (105), the light source processing system (106) and the light source are positioned in the detection area; gaps are arranged between the separation plates and the seeds conveyed on the conveying device and are used for enabling the seeds entering the monitoring area to be distributed in sequence.

4. The detection method based on the hyperspectral lower grain impurity rate detection device according to claim 3, wherein a light absorption coating is arranged on the wall surface of the detection area; the separator is made of nonmetallic materials.

5. The method according to claim 1, wherein the light source generates incident light with intensity of l _f, and the light source processing system (106) determines reflected light intensity of l _z and reflectance α of l _z/l_f according to the collected reflected light of the grain.