CN104749189A - Grain interior insect pest detection device based on multi-spectral imaging technology - Google Patents

Abstract

The invention discloses a pest detection device inside grain grains based on multi-spectral imaging technology, which comprises a grain grain sampling device and a machine vision processing device installed on a grain grain transmission device; the grain grain sampling device comprises a screw conveyor (1), one end of the screw conveyor (1) is equipped with a first motor (2), and a mesh screen (3) is installed at the outlet of its other end; the grain conveying device includes a conveyor ( 4), the conveyor belt (5) of the conveyor (4) is equipped with a collection box (6), and one end of the conveyor (4) is equipped with a corresponding second motor (7) and a mesh screen (3) The lower photoelectric sensor (8). The invention can effectively detect the grains, integrate multiple image information, and automatically identify whether the grains are infested by pests, which has not been involved in previous documents.

Description

A device for detecting pests inside grain grains based on multispectral imaging technology

technical field

The invention discloses a method for non-destructive detection of pests inside grain grains, in particular to a detection device for pests inside grain grains based on multi-spectral imaging technology.

Background technique

Food is the basic material condition for human survival and development, and food security is closely related to human survival, national security and social development. Wheat is one of the main food crops in my country. It is not only the main food resource, but also an important industrial raw material. However, the growth and reproduction of warehouse insects can cause huge losses in seed storage. According to the survey data of the United Nations Food and Agriculture Organization, the world's annual The loss of moldy grain is 3%, and the loss of insect damage is 10%. The grain damaged by stored grain pests can be eaten by more than 200 million people a year. In addition, when warehouse insects carry out harmful activities in the warehouse, the temperature and moisture of the seeds will increase, which seriously threatens the safety of seed storage. Therefore, it is very important to do a good job of pest protection in wheat granaries.

Wheat has no hard outer shell, the cortex is thin and the tissue is soft, and it has poor insect resistance and is susceptible to insect pests. Rice weevils and grain beetles are moth-eating pests that do serious harm to wheat grains. Their adults gnaw on grains outside. The females lay eggs in wheat grains and secrete mucus randomly to seal the hole. The eggs laid develop inside the endosperm, and the larvae The damage of moth is always in the grain grains for a long time, and it is very concealed and difficult to detect from the outside. The harm of pests to wheat is not only weight loss, but also insect feces and their excrement will pollute the grain. When the infection and activity of pests are serious, it will even cause wheat to heat up and mildew, further leading to a decline in wheat quality. It can be seen that the damage caused by pests inside grain grains is quite large, and detection is relatively difficult. Therefore, how to find, and accurately and automatically detect the pests inside grain grains is very important.

The detection of pests inside grain grains has been a research hotspot in the field of grain insect detection for more than ten years. Traditional detection methods for pests inside grains include sensory inspection, Burris funnel method, pest fragment inspection method, suspension method, etc. These methods have one or more disadvantages, such as too subjective, complicated and cumbersome, inaccurate, time-consuming, destructive, etc. In recent years, some new detection methods have been used in the detection of pests in grains, such as CT imaging, electrical conductivity, electronic nose, thermal imaging, etc.

Toews et al. (2006) studied the application of CT imaging method to analyze the infestation of pests inside the grain. This method needs to add vegetable oil to the grain sample to increase the contrast of the image, and the equipment cost is too high. Pearson et al. (2007) proposed to use the conductivity method to detect pests inside grain grains. This method needs to crush grain grains and cannot detect grain grains infested by larvae. In 2007, Zhang Hongmei et al. used electronic nose sensor arrays to judge whether grains were infested by pests. The sample preparation and sampling time of this method was too long. In 2008, Manickavasagan et al. used a thermal imager to judge whether each wheat grain was infested by pests. The detection process of this method was too cumbersome and the sample preparation time was too long. Currently, most of these new detection methods cannot detect low-level infestation and do not have the potential for automatic detection.

The three detection methods of soft X-ray imaging, near-infrared spectroscopy and near-infrared imaging can determine the physical and chemical changes in grains caused by pests, and can be used for automatic detection of grains. In 2006, Fornala et al. analyzed the scanned images of insect-infected grain imaging film, and the prediction accuracy rate was over 90%. Soft X-ray imaging can identify grain grains infested by early grain insects, but the radiation may pose health risks to humans. Karunakaran et al. (2005) concluded that near-infrared spectroscopy is not reliable for the quantification of insect-containing grains when the infection level is relatively low. Zhang Hongtao and others proposed to use near-infrared imaging to detect the infestation of pests inside grain grains. The equipment cost is high, and the recognition rate needs to be further improved.

To sum up, there are endless detection methods for stored grain pests at home and abroad. The traditional detection methods are cumbersome to operate, poor in accuracy, and destructive; the current newer methods are not very effective in identifying the situation where the degree of infestation is relatively low. . In view of the above reasons, there is a need for a detection method that can quickly, accurately, and non-destructively detect pests inside grain grains in the early stage of pest infestation, so as to guide the modern storage management of wheat and reduce the damage caused by failure to detect and deal with them in time. economic loss.

Contents of the invention

The technical problem to be solved by the present invention is: Aiming at the defects of the existing detection technology, a kind of internal pest detection device based on multi-spectral imaging technology is proposed, which can automatically obtain the effective information of multi-spectral images, construct the optimal feature space, establish The intelligent classification model distinguishes whether there are pests inside the grains, and realizes real-time, accurate and non-destructive automatic detection of pests inside the grains.

In order to solve the existing problems in the background technology, it includes a grain sampling device and a machine vision processing device installed on the grain conveying device; the grain sampling device includes a screw conveyor 1, and one end of the screw conveyor 1 A first motor 2 is installed, and a mesh screen 3 is installed at the discharge port at the other end; the grain conveying device includes a conveyor 4, and a collection box 6 is installed on the conveyor belt 5 of the conveyor 4, and the One end of the conveyor 4 is equipped with a corresponding second motor 7 and a photoelectric sensor 8 positioned below the mesh screen 3, and its other end is provided with a collection box 9; the machine vision processing device includes a multispectral camera positioned above the conveyor belt 5 10. A light source 11 is installed on both sides of the multispectral camera 10, and its information output terminal is connected to the computer 12 through a wire; the control terminals of the first motor 2 and the second motor 7 are respectively connected to the control box 13 through a wire, so that The information output ends of the computer 12 and the photoelectric sensor 8 are connected to the information input ends of the control box 13 respectively.

One end of the bottom of the collection box 6 is fixed on the conveyor belt 5 through a hinge, and the other end is equipped with a foot 14 that can be horizontally supported on the conveyor belt 5 .

Due to the adoption of the above technical solutions, the present invention has the following beneficial effects: (1) It can effectively detect grains, integrate multiple image information, and automatically identify whether grains are infested by pests, which has not been found in previous documents Involved; (2) Through image fusion and principal component analysis, the effective feature information of grain grains was extracted, and a classification model was established, so that the real-time classification accuracy rate of whether grain grains were infested by pests reached more than 95%; (3) The automatic grain transmission and layered sampling device is organically combined with the multi-spectral imaging device to realize the accurate judgment of whether the grain is infected by pests and improve the detection efficiency of pests inside the grain.

Description of drawings

In order to illustrate the present invention more clearly, the embodiments will be briefly introduced below in conjunction with the accompanying drawings.

Fig. 1 is a schematic structural view of the present invention.

Detailed ways

In order to make the technical means, creative features, goals and effects achieved by the present invention easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

Referring to Fig. 1, the present embodiment is realized by adopting the following technical scheme, which includes a grain sampling device and a machine vision processing device installed on a grain conveying device; the grain sampling device includes a screw conveyor 1, which One end of the screw conveyor 1 is equipped with a first motor 2, and a mesh screen 3 is installed at the outlet of the other end; the grain conveying device includes a conveyor 4, and the conveyor belt 5 of the conveyor 4 is equipped with Acquisition box 6, one end of described conveyer 4 is equipped with corresponding second motor 7 and the photoelectric sensor 8 that is positioned at mesh mouth sieve 3 below, and its other end is provided with collection box 9; 5, the multispectral camera 10 above, the both sides of this multispectral camera 10 are equipped with light source 11, its information output end is connected with computer 12 by wire; The control end of described first motor 2 and second motor 7 is connected with by wire respectively The control box 13 is connected, and the information output ends of the computer 12 and the photoelectric sensor 8 are respectively connected with the information input ends of the control box 13 .

One end of the bottom of the collection box 6 is fixed on the conveyor belt 5 through a hinge, and the other end is equipped with a foot 14 that can be horizontally supported on the conveyor belt 5 .

Grains slowly enter the conveying pipeline through the feeding port of the screw conveyor 1, and then the first motor 2 drives the screw mandrel to rotate at a certain speed, and the grains slide forward along the direction of the helical core. The tail of the screw conveyor 1 is provided with a discharge port, and the mesh sieve 3 is located directly below the discharge port, and each hole of the mesh sieve 3 is slightly larger than the diameter of two wheat grains. Thereby the first motor 2 controls the rotational speed of the bearing, and the inclination of the spiral blade set on the spiral core and the distance between the blades control the quantity of grains, and after the grain grains that fall pass through the mesh screen 3, the grains that fall into The grains have just realized the purpose of quantitative, non-destructive and single-layer tiling falling into the collection box 6, and the control box 13 is respectively connected with the photoelectric sensor 8, the computer 12, the first motor 2 and the second motor 7 through the data lines, and the computer 12 The trigger signal from the photoelectric sensor 8 is received through the control box 13 , and control commands are issued to the first motor 2 and the second motor 7 through the control box 13 . The conveyor belt 5 preferably adopts a double chain transmission mechanism, driven by the first motor 2 to run. The right end of collection box 6 is hinged on the double chain, and the left end is a free end. The collection boxes 6 are equally spaced on the conveyor belt 5. When the leftmost collection box 6 in Fig. When the box 6 is in place, the control box 13 sends a signal that the second motor 7 stops and the first motor 2 starts. After a delay of 3 seconds, the first motor 2 stops and starts the multispectral camera 10 to take pictures. When collection box 6 moved to conveyer belt 5 rightmost limit positions, collection box 6 turned over automatically under the effect of gravity, toppled over the grain grain in collection box 6, was collected by the collection box 9 that is positioned at its just below.

The light sources 11 installed on both sides of the multi-spectral camera 10 provide uniform diffuse reflection light for the grains in the collection box 6, and can clearly take multi-spectral images of the grains in the collection box 6, and finally the captured images are transmitted to The computer 12 performs processing.

Below in conjunction with accompanying drawing, the using method and principle of technical solution part in this concrete implementation are further elaborated:

During work, at first determine the rotational speed of the first motor 2 and the second motor 7, adjust the intensity of the light source 11 and the exposure time of the multispectral camera 10, to collect clear grain images. The grain sample is poured into the feeding port, and the conveyor belt 5 runs under the drive of the second motor 7 . After the photoelectric sensor 8 detects that the collection box 6 is in place, the computer 12 sends a control command to the control box 13 and starts the first motor 2 to drive the screw mandrel to rotate, and drive the screw blade to rotate, and the grains are delivered to the discharge port. Quantitative grains fall into the collection box 6 evenly through the mesh sieve 3 . After a delay of 3 seconds, the computer 12 sends a stop instruction to the first motor 2 through the control box 13 and sends an image acquisition instruction to the multispectral camera 10 . After the image acquisition is completed, the conveyor belt 5 moves on. Image processing is performed on the collected multispectral images, and the image fusion effect is evaluated by principal component analysis, and the optimal image combination is obtained as IR-R, G. Analyze the gray value range of the grain-infested area in the optimal image combination, extract its gray histogram statistical features, texture features (standard deviation, smoothness, third-order moment) parameters, and use the intelligent recognition model to determine whether the grain is Infested, and the statistics of the infested grains were carried out. After the exposure time of the multispectral camera 10, the computer controls the second motor 7 to run again. When the collection box moves to the rightmost limit position of the conveyor belt 5, the collection box performs a button-turning action, and the grain grains in the collection box are dumped. The collection box 9 below is collected for the next detection.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (2)

1., based on a grain intragranular portion injurious insect detector for multi-optical spectrum imaging technology, it is characterized in that it comprises and be arranged on grain grain sampler on grain grain transmitting device and machine vision treating apparatus; Described grain grain sampler comprises worm conveyor (1), and one end of this worm conveyor (1) is provided with the first motor (2), and the discharge outlet of its other end is provided with network interface sieve (3); Described grain grain transmitting device comprises conveyer (4), the travelling belt (5) of this conveyer (4) is provided with collecting cassette (6), one end of described conveyer (4) is provided with corresponding second motor (7) and is positioned at the photoelectric sensor (8) of network interface sieve (3) below, and its other end is provided with disposable box (9); Described machine vision treating apparatus comprises the multispectral camera (10) being positioned at travelling belt (5) top, the both sides of this multispectral camera (10) are provided with light source (11), and its information output is connected with computing machine (12) by wire; Described first motor (2) is connected with control box (13) respectively by wire with the control end of the second motor (7), and described computing machine (12) is connected with the information input terminal of control box (13) respectively with the information output of photoelectric sensor (8).

2. a kind of grain intragranular portion injurious insect detector based on multi-optical spectrum imaging technology according to claim 1, it is characterized in that one end of described collecting cassette (6) bottom is fixed on travelling belt (5) by hinge, its other end is provided with can the leg (14) of horizontal support on travelling belt (5).