TWI720842B - Intelligent bamboo shoot cultivation and harvesting monitoring system and method - Google Patents

Abstract

An intelligent bamboo shoot cultivation and harvesting monitoring system, which implements an intelligent bamboo shoot cultivation and harvesting monitoring method, includes: arranging a camera module on a mobile device to obtain an image data of a bamboo shoot; and capturing an image of the mobile device The module captures the image data and sends it to an image judging module of the mobile device; the image judging module on the mobile device real-time judges that a judgment result of the image data is to be processed or does not need to be processed according to an image deep learning model , The image judging module judges whether the judging result is to be harvested or to be covered with soil based on a long lifetime and a preset period of time. In this way, the judgment result of the image data can be directly obtained on the mobile device, and the image data does not need to be sent back to the server for judgment every time.

Description

Intelligent bamboo shoot cultivation and harvesting monitoring system and method

The invention relates to a bamboo shoot cultivation and harvesting monitoring system and method, in particular to an intelligent bamboo shoot cultivation and harvesting monitoring system and method.

When growing crops, farmers often need to visit frequently. However, with the aging of farmers, the increase in personnel costs caused by rising wages, and the low willingness of young people to farm, farmers are gradually using electronic monitoring systems to replace traditional labor costs. Manpower inspection.

For example, the Republic of China Patent Announcement No. I662505 provides a fungus growth image monitoring system, which is suitable for indoor fungus cultivation fields. The indoor fungus cultivation field includes at least a culture rack and the fungi carried on the culture rack. The fungus growth image monitoring system includes an aerial camera. With the management host. The aerial camera has a positioning module and an image capturing module. The positioning module has a plurality of ultrasonic sensing elements for positioning. The image capturing module is used to shoot fungi to generate image data. The management host communicates with the aerial camera, and is adapted to drive the aerial camera to move along a preset path according to the scheduled trajectory database and the positioning data of these ultrasonic sensing components.

However, in the aforementioned patent case, after the aerial camera generates the image data, the image data must be sent to the management host to generate the judgment data in the management host, and the judgment data cannot be generated on the aerial camera. In addition, the user cannot directly know from the judgment data what action should be taken and whether it can be harvested, and there are still many inconveniences in use.

In this regard, the present inventor proposes an intelligent bamboo shoot cultivation and harvesting monitoring system for shooting a bamboo shoot. The intelligent bamboo shoot cultivation and harvesting monitoring system includes: a server end; a mobile device including a single board computer. The single-board computer signal is connected to the server. The single-board computer has an image capture module, an image judgment module, and a GPS module. The image judgment module is a graphics processor, and the image judgment module stores a data Database, the database is an image deep learning model; and a camera module, which is set on the mobile device and signally connected to the image capture module; when the camera module obtains an image data of the bamboo shoot, the image The judgment module on the mobile device judges in real time whether a judgment result of the image data is to be processed or does not need to be processed according to the database, and returns the judgment result to the server; when the image data conforms to the database In the case of an emerging image, the image determination module determines that the determination result is pending processing. When the image data matches an unemerged image in the database, the image determination module determines that the determination result does not require processing; the emerging image is The bamboo shoot has a head, and the non-head image means that the bamboo shoot has no head; when the image data matches the head image in the database, the image judgment module compares the lifetime of the GPS module with a preset cycle time; When the growth time matches the preset cycle time, the image determination module determines that the determination result is to be harvested; when the growth time does not match the preset cycle time, the image determination module determines the The judgment result is to be covered with soil. In the embodiment of the present invention, the mobile device is a self-propelled vehicle.

Furthermore, the server side includes an image storage module.

Furthermore, a user terminal signal is connected to the server terminal, and the user terminal includes an image viewing module.

Further, the single board computer includes a battery detection module.

Further, the camera module includes an infrared light camera and a visible light camera.

The inventor further proposes a smart bamboo shoot cultivation and harvesting monitoring method, which includes: arranging a camera module on a mobile device to obtain an image data of a bamboo shoot; an image capturing module of the mobile device capturing the image The data is sent to an image judgment module of the mobile device; the image judgment module real-timely judges a judgment result of the image data on the mobile device according to a database as to be processed or does not need to be processed, when the image data conforms to the data When an emerging image of the database is generated, the image determination module determines that the determination result is to be processed, and when the image data matches a non-emergent image of the database, the image determination module determines that the determination result does not need to be processed; The image shows that the bamboo shoot has a head, and the non-head image is that the bamboo shoot has no head; when the image data matches the head image of the database, the image judgment module compares the lifetime of a GPS module with a preset period Time; when the growth time matches the preset cycle time, the image judging module judges that the judgment result is to be harvested; when the growth time does not match the preset cycle time, the image judging module It is judged that the judgment result is to be covered with soil.

Further, before the image judgment module judges the judgment result in real time, a server side establishes the database based on the plural photos of the bamboo shoots, and the server side transmits the database to the image judgment module.

Further, the database is an image deep learning model. When the image data does not match the emergent image and does not meet the non-emergence image, the image determination module sends the image data to the server, and the image data is After marking, the server side updates the image deep learning model according to the image data.

According to the above technical features, the following effects can be achieved:

1. The judgment result of the image data can be directly obtained on the mobile device, without the need to send the image data back to the server to make the judgment every time, and judge whether the bamboo shoots have emerged or not. It needs to be processed, and then the growth time of the GPS module is used to determine what processing to do, which greatly reduces the burden on the user.

2. The infrared light camera and the visible light camera take the image data together, so there is no need to worry about the accuracy of the night or bad weather.

3. The battery detection module can prompt the user first when the power is lower than the set threshold, so as to avoid the sudden loss of power during the use of the mobile device, which will cause the user to be troubled.

4. When the judgment result cannot be obtained directly on the mobile device, the image judgment module will return the image data to the server, and the image deep learning model can be updated.

1: Self-propelled car

11: Photography module

111: Infrared camera

112: Visible light camera

12: Ultrasonic module

13: Lithium battery

14: Single board computer

141: Image capture module

142: Image Judgment Module

143: Battery detection module

144: GPS module

2: server side

21: Network module

22: Image storage module

23: report chart module

24: Message notification module

3: User side

31: Image viewing module

32: Self-propelled car control module

33: Message receiving module

[The first figure] is a schematic diagram of the implementation of the embodiment of the present invention.

[The second figure] is a system block diagram of an embodiment of the present invention.

[Third Figure] is the first flow chart of the embodiment of the present invention.

[Fourth Figure] is the second flowchart of the embodiment of the present invention.

Based on the above technical features, the main effects of the intelligent bamboo shoot cultivation and harvesting monitoring system and method of the present invention will be clearly presented in the following embodiments.

Please refer to the first and second figures, an embodiment of the intelligent bamboo shoot cultivation and harvesting monitoring system of the present invention is disclosed, including: a self-propelled vehicle (1), a server end (2), and a user end (3) ).

The self-propelled vehicle (1) is provided with a photographing module (11), an ultrasonic module (12), a lithium battery (13) and a single board computer (14). The single-board computer (14) is connected to the server end (2) with signals. The single-board computer (14) has 5T computing power. The single-board computer (14) includes an image capture module (141) and an image The judgment module (142), a battery detection module (143) and a GPS module (144), the image judgment module (142) For the graphics processor. The photographing module (11) includes an infrared light camera (111) and a visible light camera (112), and the photographing module (11) is signally connected to the image capturing module (141). With the infrared light camera (111) and the visible light camera (112), there is no need to worry that night or bad weather will reduce the accuracy of the image data. The infrared light camera (111) and the visible light camera (112) can be used again. A USB Video Class (UVC) is connected to a computer to transmit the images of the infrared light camera (111) and the visible light camera (112) to the computer.

The server end (2) is connected to the single-board computer (14) via a wireless network signal. The server end (2) includes a network module (21), an image storage module (22), and a Report chart module (23) and a message notification module (24).

The user end (3) is signaled to the server end (2), and the user end (3) includes an image viewing module (31), a self-propelled vehicle control module (32) and a message receiving module (33).

Please refer to the third figure with the second figure. The server (2) creates a database based on the plural photos of a bamboo shoot. The database is an image deep learning model in the embodiment of the present invention. The photo is taken manually in advance, and an emerging image or a non-emerging image of the photo is marked one by one. The rising image is that the bamboo shoot has a rising head, and the non-rising image is that the bamboo shoot has no rising head. More specifically, it is judged as the rising image or the non-rising image based on whether there is a shoot tip of the bamboo shoot in the photo. The server (2) then sends the image deep learning model to the image judgment module (142).

Please refer to the fourth figure, and please match the second figure. After the camera module (11) is installed on the self-propelled vehicle (1), the self-propelled vehicle (1) can be operated by the single board computer (14) The GPS module (144) of the GPS module (144) plans the route, the self-propelled car control module (32) of the user terminal (3) operates the route, or the ultrasonic module (12) automatically avoids obstacles Move. After the self-propelled vehicle (1) moves, the photographing module (11) photographs the location of the bamboo shoot to obtain an image data of the bamboo shoot. The image data is in a video format. The location can be stored in the GPS module (144) in advance. The image capturing module (141) captures the image data by edge detection, and sends the image data to the image determining module (142), and the image determining module (142) is in the self-propelled vehicle (1). ) Directly based on the image deep learning model to determine a judgment result of the image data in real time, and it is not necessary to return the image data to the server (2) every time to obtain the judgment result. When the image data matches the emerging image of the image deep learning model, the image determination module (142) determines that the determination result is to be processed. When the image data matches the unemerged image of the image deep learning model, the image determination module (142) determines that the result of the determination is to be processed. The image judgment module (142) judges that the judgment result is no need to process.

When the image data matches the emergence image of the image deep learning model, the image determination module (142) compares the lifetime long time stored by the GPS module (144) with a preset period time, and the growth time is the bamboo shoot The actual growth time can be calculated from the planting time of the bamboo shoots also stored in the GPS module (144) and the time when the photographing module (11) obtains the image data; when the growth time and the When the preset cycle time matches, the image judging module (142) judges that the judging result is to be harvested; when the growth time does not match the preset cycle time, the image judging module (142) judges the The judgment result is to be covered with soil. For example, if the preset cycle time of the bamboo shoots is between N-3 days and N+3 days, when the growth time of the GPS module (144) is N-1 days, it conforms to the preset Cycle time, the image judgment module (142) will judge the judgment result to be harvested; and when the growth time of the GPS module (144) is N-5 days, it does not meet the preset cycle time, The image judgment module (142) judges that the judgment result is to be covered with soil. By judging whether the emerging image and the unemerged image are to be processed or need not be processed, and then judging whether it is to be harvested or to be covered by the growth time, the appropriate processing method for the bamboo shoot can be quickly determined, which greatly reduces the user Burden.

In addition to the identification of the rising image and the non-rising image of the bamboo shoot, the image judging module (142) can also judge the soil moisture according to the soil color. In more detail, you can choose a fertilizer first, The photographing module (11) is used to photograph soils with different humidity under the fertilizer, so as to establish the image deep learning model of soil moisture on the server side (2), and send it to the image judgment module (142), Then the image judging module (142) can compare the image deep learning model of soil moisture with the image data to obtain the soil moisture.

After the image judgment module (142) judges the judgment result, the image judgment module (142) compresses the image data to the H.264 standard, and sends it to the server via a virtual private network (Virtual Private Network, VPN) The server end (2) to store the image data to the image storage module (22), the server end (2) and also send the image data to the user end (3) through the VPN for viewing the image The module (31) is for a user to watch. With the VPN, the streaming of the image data can be prevented from being interfered by others. At the same time, the image judgment module (142) sends the judgment result to the message receiving module (33) of the user end (3) via the message notification module (24) of the server side (2). The user watched.

When the image data does not match the emerging image and does not match the non-emerging image, the image determination module (142) sends the image data to the server (2), and marks the image data as After the emerging image or the non-emerging image, the server end (2) updates the image deep learning model according to the image data, and stores the image data in the image storage module (22).

When the user processes the bamboo shoots, the image data will change. For example, the user appears in the image data, the position of the self-propelled vehicle (1) changes or the position of fallen leaves, etc., the image judgment model The group (142) judges whether the user has to process the bamboo shoots. The report chart module (23) of the server (2) then creates a report chart of the judgment result and the actual processing situation based on the image data to modify the image deep learning model. For example, the judgment result is When there is no need to process, the user processes the bamboo shoots, which means that the image deep learning model needs to be revised to improve the accuracy of the image deep learning model.

The power of the self-propelled vehicle (1) is supplied by the lithium battery (13). When the power of the lithium battery (13) is lower than a set threshold, the battery detection module (143) will pass through the server ( 2) The message notification module (24) sends a prompt message to the message receiving module (33) of the user terminal (3) to notify the user that the lithium battery (13) needs to be replaced or the The lithium battery (13) is charged to prevent the self-propelled vehicle (1) from suddenly running out of power during use, which may cause trouble to the user.

Based on the description of the above embodiments, when one can fully understand the operation and use of the present invention and the effects of the present invention, but the above embodiments are only the preferred embodiments of the present invention, and the implementation of the present invention cannot be limited by this. The scope, that is, simple equivalent changes and modifications made according to the scope of the patent application of the present invention and the content of the description of the invention, are all within the scope of the present invention.

1: Self-propelled car

11: Photography module

111: Infrared camera

112: Visible light camera

12: Ultrasonic module

Claims (8)

A smart bamboo shoot cultivation and harvesting monitoring system for shooting a bamboo shoot. The smart bamboo shoot cultivation and harvesting monitoring system includes: a server end; a mobile device including a single board computer, the single board computer signal is connected to the servo On the device side, the single-board computer has an image capture module, an image judgment module, and a GPS module. The image judgment module is a graphics processor. The image judgment module stores a database, and the database is a Image deep learning model; and a photographing module, which is arranged on the mobile device and signally connected to the image capturing module; when the photographing module obtains an image data of the bamboo shoot, the image determining module is on the mobile device According to the database, it determines in real time that a judgment result of the image data is to be processed or does not need to be processed, and returns the judgment result to the server; when the image data matches an emerging image of the database, the image The judgment module judges that the judgment result is to be processed. When the image data matches a non-emerging image in the database, the image judgment module judges that the judgment result does not need to be processed; The emergence image is that the bamboo shoot has no emergence; when the image data matches the emergence image of the database, the image determination module compares the lifetime of the GPS module with a predetermined period time; when the growth time is compared with the predetermined period When the cycle time is consistent, the image judgment module judges that the judgment result is to be harvested; when the growth time does not conform to the preset cycle time, the image judgment module judges that the judgment result is to be covered with soil. According to the intelligent bamboo shoot cultivation and harvesting monitoring system described in claim 1, further, the server side includes an image storage module. According to the intelligent bamboo shoot cultivation and harvesting monitoring system described in claim 1, further, a user end signal is connected to the server end, and the user end includes an image viewing module. According to the intelligent bamboo shoot cultivation and harvesting monitoring system described in claim 1, further, the single board computer includes a battery detection module. According to the intelligent bamboo shoot cultivation and harvesting monitoring system described in claim 1, further, the photographing module includes an infrared light camera and a visible light camera. An intelligent bamboo shoot cultivation and harvesting monitoring method includes: setting a camera module on a mobile device to obtain an image data of a bamboo shoot; an image capturing module of the mobile device captures the image data and sends it to the mobile device An image judging module of the device; the image judging module on the mobile device real-time judges whether a judgment result of the image data is to be processed or does not need to be processed according to a database, when the image data matches an emerging image of the database When the image judgment module judges that the judgment result is to be processed, when the image data matches an unemerged image in the database, the image judgment module judges that the judgment result does not need to be processed; the emergence image is that the bamboo shoot has When the image data matches the rising image of the database, the image judging module compares the lifetime of a GPS module with a preset cycle time; when the image is growing When the time matches the preset cycle time, the image judgment module judges that the judgment result is to be harvested; when the growth time does not match the preset cycle time, the image judgment module judges that the judgment result is To be covered with soil. According to the intelligent bamboo shoot cultivation and harvest monitoring method described in claim 6, further, before the image judgment module judges the judgment result in real time, a server side establishes the database according to the plural photos of the bamboo shoots, and the server side And send the database to the image judgment module. For example, the intelligent bamboo shoot cultivation and harvesting monitoring method described in claim 7, further, the database is an image deep learning model, when the image data does not meet the emerging image and does not meet the non-emergence image, the image judgment model The group sends the image data to the server side. After the image data is marked, the server side updates the image deep learning model according to the image data.

Images