阅读说明：本技术 一种基于机器视觉的稻飞虱监测装置及监测方法 (Rice planthopper monitoring device and monitoring method based on machine vision ) 是由 刘双喜 张正辉 刘印增 白壮壮 黄信诚 胡宪亮 高发瑞 于 2021-07-15 设计创作，主要内容包括：本发明涉及一种基于机器视觉的稻飞虱监测装置及监测方法,其中的一种基于机器视觉的稻飞虱监测装置,包括收集盒,以及设置在收集盒内的黄色诱捕粘板和图像采集装置,图像采集装置的图像采集端朝向黄色诱捕粘板,收集盒的侧壁上设有隔离栅栏,隔离栅栏的栅栏孔尺寸大于飞虱尺寸、小于飞蛾尺寸。本发明通过黄色诱捕粘板、蓝色诱捕灯和隔离栅栏的设置,使飞虱聚集在飞虱收集腔,便于图像采集装置对飞虱情况进行图像采集,图像采集装置采集的图像通过无线传输传至计算机,利用已训练好的卷积神经网络模型对图像进行识别,从而可以及时关注稻飞虱的数量、种类,以便及时制定调整防治策略,减少虫害对稻田的危害,提高水稻的产量。(The invention relates to a rice planthopper monitoring device and a monitoring method based on machine vision, wherein the rice planthopper monitoring device based on machine vision comprises a collecting box, a yellow trapping sticky plate and an image acquisition device, wherein the yellow trapping sticky plate and the image acquisition device are arranged in the collecting box, the image acquisition end of the image acquisition device faces the yellow trapping sticky plate, a barrier is arranged on the side wall of the collecting box, and the size of a barrier hole of the barrier is larger than that of a planthopper and smaller than that of a moth. According to the rice planthopper collecting device, the yellow trapping sticky plates, the blue trapping lamps and the isolation fences are arranged, so that planthoppers are gathered in the planthopper collecting cavity, an image collecting device can conveniently collect images of the planthoppers, the images collected by the image collecting device are transmitted to a computer through wireless transmission, and the images are identified by using a trained convolutional neural network model, so that the number and the types of the rice planthoppers can be paid attention to in time, a control strategy can be established and adjusted in time, damage of insect pests to a rice field is reduced, and the yield of rice is increased.)

1. The utility model provides a rice planthopper monitoring devices based on machine vision which characterized in that: the collecting box comprises a collecting box, a yellow trapping sticky plate and an image collecting device, wherein the yellow trapping sticky plate and the image collecting device are arranged in the collecting box, the image collecting end of the image collecting device faces towards the yellow trapping sticky plate, an isolation fence is arranged on the side wall of the collecting box, and the size of a fence hole of the isolation fence is larger than that of a plant hopper and smaller than that of a moth.

2. A machine vision based rice planthopper monitoring device according to claim 1, wherein: the top of collecting the box sets firmly the seal box, and image acquisition device is located the seal box, is equipped with the shooting hole with image acquisition device's image acquisition end adaptation on the bottom plate of seal box, yellow is traped the sticky board and is fixed in the bottom of collecting the box, forms the planthopper between the bottom plate of isolation fence, seal box and the bottom plate of collecting the box and collects the chamber.

3. A machine vision based rice planthopper monitoring device according to claim 2, wherein: the lower surface of the bottom plate of the sealing box is fixed with a blue trapping lamp which is arranged annularly.

4. A machine vision based rice planthopper monitoring device according to claim 3, wherein: the lower surface of the bottom plate of the sealing box is also fixed with a lighting lamp which is arranged annularly, and the lighting lamp is positioned on the annular inner side of the blue trapping lamp.

5. A machine vision based rice planthopper monitoring device according to claim 4, wherein: and a control device electrically connected with the illuminating lamp and the image acquisition device and a storage battery for supplying power to the blue trapping lamp, the illuminating lamp, the image acquisition device and the control device are also arranged in the sealing box.

6. A machine vision based rice planthopper monitoring device according to claim 5, wherein: and a solar power supply device electrically connected with the storage battery is fixedly arranged above the collecting box.

7. A machine vision based rice planthopper monitoring device according to claim 2, wherein: the collecting box also comprises a telescopic vertical rod, and the upper end of the vertical rod is fixedly connected with the bottom plate of the collecting box.

8. A machine vision based rice planthopper monitoring apparatus according to claim 7, wherein: the pole setting includes the interior pole with the rigid coupling of collection box to and the outer tube of locating interior pole lower extreme, have along the adjusting bolt of horizontal top to interior pole through threaded connection on the pipe wall of outer tube.

9. A rice planthopper monitoring method based on machine vision is characterized by comprising the following steps:

(1) vertically fixing the upright rod in a rice field, and enabling the collection box to be 10-20 cm higher than the plant leaf surface;

(2) turning on a blue trapping lamp and keeping the blue trapping lamp normally on, introducing the plant hoppers to a plant hopper collecting cavity between the image acquisition device and the yellow trapping sticky plate through the yellow trapping sticky plate and the blue trapping lamp, and preventing the plant hoppers from entering the plant hopper collecting cavity by using an isolation fence;

(3) every certain time, carrying out image acquisition on the planthopper collection cavity through an image acquisition device, improving the brightness during image acquisition by using an illuminating lamp, starting the illuminating lamp before each image acquisition, and turning off the illuminating lamp after each image acquisition;

(4) the image collected by the image collecting device is wirelessly transmitted to a computer, and the collected image is identified and segmented by an image identification system carried in the computer, so that the type and the outbreak condition of the insect pest are determined.

Technical Field

The invention relates to the technical field of rice insect pest identification, in particular to a rice planthopper monitoring device and method based on machine vision.

Background

With the continuous development of agriculture, smart agriculture becomes a mainstream development direction, insect pest identification technology is developed, the traditional insect pest monitoring device has a good detection effect on the insect pest situation of the moths, but the traditional detection device cannot effectively capture and identify the small insect pest of the planthoppers, and no device specially used for monitoring the rice planthoppers exists at present, so that the outbreak time of the rice planthoppers cannot be predicted in time, and a great amount of loss is caused to growers.

Disclosure of Invention

Aiming at the defects of the prior art, the invention provides the rice planthopper monitoring device and the monitoring method based on machine vision, so that the rice planthopper is monitored, and a basis is provided for the treatment and prevention of insect pest outbreak. Machine vision means that an image acquisition device is arranged on a mechanical device or equipment to enable the machine to have vision.

The invention is realized by the following technical scheme, and provides a rice planthopper monitoring device based on machine vision, which comprises a collecting box, and a yellow trapping sticky plate and an image acquisition device which are arranged in the collecting box, wherein the image acquisition end of the image acquisition device faces the yellow trapping sticky plate, a barrier is arranged on the side wall of the collecting box, and the size of a barrier hole of the barrier is larger than that of a planthopper and smaller than that of a moth.

The monitoring device of this scheme is when using, traps the sticky plate through yellow and introduces the planthopper in collecting the box to prevent the moth to get into through isolation fence and collect the box, in order to avoid causing the influence to the monitoring effect, the planthopper is introduced and forms the gathering in the collection box after, carries out image acquisition to the planthopper condition of collecting the box through image acquisition device, provides the reference for the insect pest analysis.

Preferably, the top of the collecting box is fixedly provided with a sealing box, the image acquisition device is positioned in the sealing box, a shooting hole matched with the image acquisition end of the image acquisition device is arranged on a bottom plate of the sealing box, the yellow trapping sticky plate is fixed at the bottom of the collecting box, and a planthopper collecting cavity is formed among the isolation fence, the bottom plate of the sealing box and the bottom plate of the collecting box. According to the optimized scheme, the sealing box is arranged to protect the image acquisition device, and the bottom plate of the sealing box is used for blocking the plant hoppers, so that the plant hoppers are prevented from flying above the shooting hole, and the dead zone of image acquisition is favorably reduced; and the sealing box is used for shielding the planthopper collecting cavity, so that the washing of rainwater to the planthopper is reduced.

Preferably, the blue trapping lamps are fixed on the lower surface of the bottom plate of the sealing box and are arranged annularly. This optimization scheme further improves the induction effect to the planthopper through setting up blue trapping lamp, sets up blue trapping lamp into the annular, has increased the illumination area, does benefit to the planthopper simultaneously and locates the gathering at annular trapping lamp.

As optimization, the lower surface of the bottom plate of the sealing box is also fixedly provided with an illuminating lamp which is arranged annularly, and the illuminating lamp is positioned on the annular inner side of the blue trapping lamp. The optimized scheme increases the brightness of the collected image by arranging the illuminating lamp so as to improve the quality of the collected image.

And as optimization, a control device electrically connected with the illuminating lamp and the image acquisition device and a storage battery for supplying power to the blue trapping lamp, the illuminating lamp, the image acquisition device and the control device are further arranged in the sealing box. This optimization scheme is through setting up controlling means, and the collection work of opening and close and image acquisition device to the lighting lamp is conveniently controlled, improves automatic level, provides the electric energy for each power consumption device through setting up the battery for this monitoring devices can use in the paddy field for a long time, has also solved the problem of the inconvenient power consumption in the paddy field simultaneously, and the seal box plays the guard action to controlling means and battery.

And preferably, a solar power supply device electrically connected with the storage battery is fixedly arranged above the collecting box. The solar energy is adopted for power supply in the optimized scheme, energy conservation and environmental protection are facilitated, the solar energy-saving solar energy lighting device is suitable for being used in a rice field, and is free of shielding in the rice field, good in lighting effect and long in lighting time.

As optimization, the collecting box further comprises a telescopic vertical rod, and the upper end of the vertical rod is fixedly connected with the bottom plate of the collecting box. This optimization scheme is through setting up the pole setting, and the convenience is whole fixed in the paddy field with the device, sets up the pole setting into extending structure, and the box is collected in the adjustment of being convenient for, satisfies the use in not co-altitude rice paddy field, has improved the commonality.

Preferably, the upright rod comprises an inner rod fixedly connected with the collecting box and an outer tube sleeved at the lower end of the inner rod, and the tube wall of the outer tube is connected with an adjusting bolt transversely propping against the inner rod through threads. This pole setting extending structure who optimizes the scheme is simple, through adjusting bolt with outer tube and interior pole relatively fixed, after loosening adjusting bolt, can adjust the length that interior pole extends the outer tube, convenient operation.

The scheme also provides a rice planthopper monitoring method based on machine vision, which comprises the following steps:

1. vertically fixing the upright rod in a rice field, and enabling the collection box to be 10-20 cm higher than the plant leaf surface;

2. turning on the blue trapping lamp and keeping the blue trapping lamp normally bright, introducing the plant hoppers to a plant hopper collecting cavity between the image acquisition device and the yellow trapping sticky plate through the yellow trapping sticky plate and the blue trapping lamp, and preventing the plant hoppers from entering the plant hopper collecting cavity by utilizing the isolation fence, so that the monitoring effect of the plant hoppers on the rice plant hoppers is prevented from being influenced by the plant hoppers;

3. every certain time, the image acquisition device is used for acquiring images in the planthopper collection cavity, the illumination lamp is used for improving the brightness during image acquisition, the illumination lamp is turned on before each image acquisition, and the illumination lamp is turned off after each image acquisition, so that the electric quantity of a battery is effectively saved, and the service life is prolonged;

4. the image collected by the image collecting device is wirelessly transmitted to a computer, and the collected image is identified and segmented by an image identification system carried in the computer, so that the type and the outbreak condition of the insect pest are determined.

The invention has the beneficial effects that: through the setting of yellow trapping sticky plate, blue trapping lamp and barrier fence, make the planthopper gathering collect the chamber at the planthopper, the image acquisition device of being convenient for carries out image acquisition to the planthopper condition, the image that image acquisition device gathered passes to the computer through wireless transmission, utilizes trained convolution neural network model to discern the image to can in time pay close attention to the quantity, the kind of rice planthopper, so that in time formulate the adjustment and control strategy, reduce the harm of pest to the paddy field, improve the output of rice.

Drawings

FIG. 1 is a schematic structural view of a rice planthopper monitoring device according to the present invention;

FIG. 2 is a schematic view of the structure of the collecting box;

FIG. 3 is an isometric view of a rice planthopper monitoring device of the present invention;

shown in the figure:

1. seal box, 2, singlechip, 3, image acquisition device, 4, battery, 5, barrier fence, 6, the bottom plate of collecting the box, 7, solar cell panel, 8, blue trapping lamp, 9, light, 10, image acquisition end, 11, adjusting bolt, 12, outer tube, 13, trapping the board and place the platform.

Detailed Description

In order to clearly illustrate the technical features of the present solution, the present solution is explained below by way of specific embodiments.

As shown in fig. 1, a rice planthopper monitoring device based on machine vision comprises a collecting box, a yellow trapping sticky plate and an image acquisition device 3, wherein the yellow trapping sticky plate and the image acquisition device 3 are arranged in the collecting box, an image acquisition end 10 of the image acquisition device faces towards the yellow trapping sticky plate, the image acquisition device of the embodiment is a household wireless high-definition network camera head with a WiFi hot spot mobile phone remote monitoring camera SDK camera, and a lens of the image acquisition device faces towards the trapping plate below. The image collected by the image collecting device is transmitted to the computer, and the computer processes the image data. And a graphic processing system is carried in the computer, and the trained network recognition model is used for recognizing the acquired images so as to determine the type and outbreak condition of the insect pests. And training by using a convolutional neural network and a large number of rice planthopper experimental samples to obtain a training model, and identifying the insect pests by using the training model.

The side wall of the collecting box is provided with a separation fence 5, the size of a fence hole of the separation fence is larger than that of a plant hopper and smaller than that of a moth, and the moth is prevented from entering to influence observation. To enhance the inducing effect, the segregator barrier of this embodiment is circumferentially closed. Specifically, the isolation fence is composed of a plurality of iron wires at intervals of 5mm, is surrounded by four sides, allows rice planthoppers to freely enter, can prevent most of moth pests from entering, and prevents the monitoring effect on the rice planthoppers from being influenced.

The top of collecting the box sets firmly seal box 1, and image acquisition device is located the seal box, is equipped with the shooting hole with image acquisition device's image acquisition end adaptation on the bottom plate of seal box, platform 13 is placed at the trapboard of collecting the box bottom to yellow traping sticky board, forms the planthopper between the bottom plate of isolation fence, seal box and the bottom plate of collecting the box and collects the chamber. The lower surface of the bottom plate of the sealing box is fixed with a blue trapping lamp 8 and a lighting lamp 9, wherein the blue trapping lamp 8 and the lighting lamp 9 are annularly arranged, and the lighting lamp is positioned on the inner side of the ring of the blue trapping lamp. The blue trapping lamp and the yellow trapping sticky plate form a trapping device, the blue trapping lamp is connected with a power supply and kept normally bright to attract the planthoppers to fly around the blue trapping lamp, and the yellow trapping sticky plate is fixed on the bottom plate of the collecting box and attracts the planthoppers to stay and trap through attractant and color.

The monitoring device of this embodiment still includes the pole setting of retractable, and the upper end of pole setting and the 6 rigid couplings of bottom plate of collecting the box, 6 upper surfaces of bottom plate of collecting the box are equipped with trapboard and place platform 13. Specifically, the upright rod comprises an inner rod fixedly connected with the collecting box and an outer tube 12 sleeved at the lower end of the inner rod, and the tube wall of the outer tube is connected with an adjusting bolt 11 transversely propping the inner rod through threads.

Still be equipped with in the sealing box with light, image acquisition device electric connection's controlling means to and the battery 4 of traping lamp, light, image acquisition device and controlling means power supply for blue, collect the top of box set firmly with battery electric connection's solar power unit, solar cell panel 7 among the solar power unit installs in collecting the box upper surface, and inclination is 40 degrees. The control device of this embodiment is singlechip 2, prevents through the seal box that the rainwater from invading and damaging electronic component.

The embodiment also provides a rice planthopper monitoring method based on machine vision, which comprises the following steps:

1. vertically and fixedly inserting the upright rods into the rice field, and adjusting the height of the upright rods to enable the collecting boxes to be 10-20 cm higher than the plant leaf surfaces and enable the solar cell panels to face the south;

2. turning on the blue trapping lamp and keeping the blue trapping lamp normally bright, introducing the plant hoppers to a plant hopper collecting cavity between the image acquisition device and the yellow trapping sticky plate through the yellow trapping sticky plate and the blue trapping lamp, and preventing the plant hoppers from entering the plant hopper collecting cavity by utilizing the isolation fence, so that the monitoring effect of the plant hoppers on the rice plant hoppers is prevented from being influenced by the plant hoppers;

3. the computer is turned on, the image acquisition device is wirelessly connected, the image acquisition device is adjusted, pictures are enabled to face the insect catching plate and the definition of the pictures is adjusted, the image acquisition device and the illuminating lamp are set to work simultaneously, image acquisition is carried out in the planthopper collecting cavity through the image acquisition device every hour, the illuminating lamp is utilized to improve the brightness of image acquisition and improve the shooting effect, the image acquisition device transmits a pulse signal to the single chip microcomputer each time when a picture is to be taken, the single chip microcomputer controls the illuminating lamp to be turned on when the image is acquired, and the illuminating lamp is turned off when the image acquisition is finished, so that the electric quantity of a battery is effectively saved;

4. the image collected by the image collecting device is wirelessly transmitted to a computer, and an image recognition system carried in the computer is used for recognizing and segmenting the collected image, so that the type and the outbreak condition of the insect pest are determined, prevention and treatment measures are made in time, and specifically, the image is recognized by using a rice planthopper insect pest recognition model trained by a convolutional neural network through an image processing system.

The rice planthopper is trapped by the blue trapping lamp and the yellow trapping sticky plate, the insect pest information of the trapping plate is collected by photographing at regular time through the camera and is transmitted to the computer through wireless transmission for insect pest diagnosis, the precision is high, the data processing is fast, the development trend of the insect pest can be reflected in time, the insect pest outbreak time can be predicted more accurately, and an optimal prevention and control plan is formulated.

Of course, the above description is not limited to the above examples, and the undescribed technical features of the present invention can be implemented by or using the prior art, and will not be described herein again; the above embodiments and drawings are only for illustrating the technical solutions of the present invention and not for limiting the present invention, and the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that changes, modifications, additions or substitutions within the spirit and scope of the present invention may be made by those skilled in the art without departing from the spirit of the present invention, and shall also fall within the scope of the claims of the present invention.