Plant health evaluation apparatus and evaluation method

Abstract

An plant health evaluation apparatus and evaluation method comprise a UV light source, a RGB camera and a controller. The UV light source provides a UV light for irradiating leaves of plants. The RGB camera photographs the leaves and acquires an evaluation image. The controller connects to and controls the UV light source and the RGB camera and separates each pixel of the evaluation image into a red light component, a green light component, and a blue light component. Based on the components, the controller calculates a UV normalized difference vegetation index (uNDVI) value of each pixel and forms a uNDVI image in order to evaluate the health of plant, where the uNDVI value is (R-B) / (R+B).

Description

[Technical Field]

[0001] This invention relates to a plant health assessment device and method, particularly an assessment device and method for monitoring the chlorophyll response of plants by irradiating them with ultraviolet light. [Previous Technology]

[0002] The health of traditional outdoor vegetation is assessed by analyzing the reflectivity of plants to infrared and red light using satellites or drones. This requires removing the IR-cut in front of the camera's CCD and adding a blue filter in order to calculate the Normalized Difference Vegetation Index (NDVI), which in turn reveals the activity of chlorophyll in the plant to determine its health. Indoor crops, on the other hand, rely on commercially available plant testing instruments, ranging from small handheld devices to large equipment such as photosynthesis meters, to assess plant health. However, these instruments are very expensive and can only be used to test a single leaf, making them uneconomical for practical agricultural applications. [Summary of the Invention]

[0003] To address the aforementioned issues, this invention provides a cost-effective and simple plant health assessment device and method, which assesses plant health based on the fact that the plant's reflectance spectrum varies with external light. When a plant is irradiated with ultraviolet light, its chlorophyll emits red fluorescence, thus reflecting red light. Therefore, when photographing a leaf surface irradiated with ultraviolet light, the red light received by the camera will be stronger. Areas with stronger red light are regions where chlorophyll is activated, allowing analysis of chlorophyll distribution and activity to assess plant health, such as leaf defects or pests and diseases.

[0004] The present invention provides a plant health assessment device and assessment method. The plant health assessment device includes an ultraviolet light source, an RGB camera and a controller. The ultraviolet light source provides ultraviolet light to irradiate the plant leaves. The RGB camera captures the plant leaves and obtains an assessment image. The controller connects to the ultraviolet light source and the RGB camera and controls the ultraviolet light source and the RGB camera. The controller divides the assessment image into multiple pixels and separates each pixel into three components: red light component R, green light component G and blue light component B. Based on the separated components, the controller calculates the ultraviolet normalization difference plant index (uNDVI) value of each pixel and generates a corresponding uNDVI image to determine the plant health. The uNDVI value is (RB) / (R+B).

[0005] In one embodiment of the present invention, the controller calculates the ratio of the area of ​​all pixels with uNDVI values ​​greater than or equal to a specific value to the area of ​​the plant leaf surface to determine the plant health.

[0006] In one embodiment of the present invention, the controller converts uNDVI values ​​to grayscale values ​​through grayscale mapping processing, converts specific values ​​to grayscale specific values, and calculates the ratio of the grayscale area of ​​all pixels with grayscale values ​​greater than or equal to the grayscale specific value to the grayscale area of ​​the plant leaf surface to determine the plant health.

[0007] In one embodiment of the present invention, after obtaining the evaluation image, the controller excludes the non-plant leaf parts in the evaluation image through AI image recognition technology.

[0008] In one embodiment of the present invention, the above-mentioned plant health assessment device further includes a communication device, which is coupled to a controller to enable the controller to display a web page interface of area ratio to the user device through the communication device, and send a prompt signal corresponding to the assessed plant health. When the controller determines that the plant health is good, the prompt light is green; when the controller determines that the plant health is normal, the prompt light is yellow; when the controller determines that the plant health is abnormal, the prompt light is red.

[0009] The plant health assessment device and method of the present invention use an RGB camera with a controllable ultraviolet light source. By taking pictures of the plant leaf surface through the RGB camera, the reflectance spectrum of the leaf surface over a large area can be received, and the chlorophyll activity and distribution can be analyzed to assess the plant health. Based on the uNDVI value, the overall plant health trend can be understood, and defective leaf surfaces can be identified through the uNDVI image. It can be used for real-time monitoring of plant growth, so that environmental control conditions can be changed at any time according to the plant health status. The RGB camera can be an existing monitoring device in the plant cultivation facility, and the ultraviolet light source can also be a supplementary light source for the existing plant growth in the facility. The plant health assessment device and method of the present invention can be used to perform detection using existing equipment, so there is no need to purchase additional expensive special instruments and equipment.

Implementation Method

[0011] The invention will now be described in more detail by way of example with reference to the accompanying drawings.

[0012] Please refer to the architecture diagram of the plant health assessment device in Figure 1. The plant health assessment device 10 includes an ultraviolet light source 11, an RGB camera 12, and a controller 13. The ultraviolet light source 11 is used to irradiate the plant leaves with ultraviolet light. The RGB camera 12 is used to take pictures of the plant leaves and obtain assessment images. The controller 13 is connected to the ultraviolet light source 11 and the RGB camera 12 and is used to control the ultraviolet light source 11 and the RGB camera 12. The controller 13 divides the assessment image into multiple pixels and separates each pixel into three components: red light component R, green light component G, and blue light component B. Based on the separated components, the controller 13 calculates the ultraviolet normalization difference plant index (uNDVI) value of each pixel and generates a corresponding uNDVI image to determine the plant health. The uNDVI value is (RB) / (R+B).

[0013] In one embodiment of the present invention, the ultraviolet light source 11 may be a UVA light source with a wavelength of 320nm-400nm, and more preferably an ultraviolet light source with a wavelength of 365nm.

[0014] Please refer to Figure 2 for the application scenario diagram of the plant health assessment device method. The plant health assessment device 10 of this embodiment can be applied in plant cultivation facilities. The ultraviolet light source 11 and RGB camera 12 can be installed on the ceiling of the facility. When testing is to be performed, the ultraviolet light source 11 and RGB camera 12 can be turned on through the controller 13 to perform shooting operations, so as to monitor the plants in the facility in real time. Alternatively, the ultraviolet light source 11, RGB camera 12 and controller 13 can be installed on a mobile trolley to perform plant health testing in a specific area.

[0015] In one embodiment of the present invention, the controller 13 calculates the ratio of the area of ​​all pixels with uNDVI values ​​greater than or equal to a specific value to one of the plant leaf areas, and determines the health index of the detected plant based on the area ratio, wherein the uNDVI value and the specific value are between -1 and 1. Furthermore, those skilled in the art can adjust the size of the specific value according to the type of plant, and the present invention does not limit this specific value.

[0016] In one embodiment of the present invention, the controller 13 converts the uNDVI value to a grayscale value via grayscale mapping processing, and converts a specific value to a specific grayscale value via grayscale mapping processing. The controller 13 calculates the ratio of the grayscale area of ​​all pixels with grayscale values ​​greater than or equal to the specific grayscale value to the grayscale area of ​​the plant leaf surface to determine the plant health, wherein the grayscale value and the specific grayscale value are between 0 and 255. Furthermore, those skilled in the art can adjust the size of the specific grayscale value according to the type of plant, and the present invention does not limit this specific grayscale value.

[0017] In one embodiment of the present invention, when the controller 13 calculates that the area ratio of all pixels with uNDVI values ​​greater than or equal to a specific value to the plant leaf surface is greater than 90%, or the grayscale area ratio of all pixels with grayscale values ​​greater than or equal to a specific grayscale value to the plant leaf surface is greater than 90%, the controller 13 determines that the plant health is good. When the area ratio or grayscale area ratio is between 70% and 90%, the controller 13 determines that the plant health is normal. When the area ratio or grayscale area ratio is less than 70%, the controller 13 determines that the plant health is abnormal. Those skilled in the art can adjust the standards used by the controller 13 to determine the plant health as good, normal, or abnormal according to the type of plant; the present invention is not limited to the above values.

[0018] After obtaining the evaluation image, the controller 13 can first use AI image recognition technology to determine the plant leaf surface in the evaluation image and exclude the non-plant leaf surface in the evaluation image, and then calculate the uNDVI value to make the calculation of the area ratio more accurate.

[0019] The controller 13 performs pseudo-color rendering on the uNDVI image to display the plant health of the plant leaves, and imports an AI model to determine the types of pests and diseases in pixels with uNDVI values ​​less than a specific value. For example, green can be used to mark pixels with uNDVI values ​​greater than or equal to a specific value, and blue can be used to mark pixels with uNDVI values ​​less than a specific value. Therefore, the user can intuitively see where the leaf surface in the evaluation image is defective. In addition, the import of the AI ​​model also allows the controller 13 to more accurately identify the areas of defective leaf surfaces.

[0020] In one embodiment of the present invention, the plant health assessment device 10 further includes a communication device 14, which is coupled to a controller 13. The controller 13 provides a webpage interface through the communication device 14, allowing the user device to display the area ratio, and sends an indicator light corresponding to the assessed plant health. When the controller 13 determines the plant health is good, the indicator light is green; when the controller 13 determines the plant health is normal, the indicator light is yellow; and when the controller 13 determines the plant health is abnormal, the indicator light is red. The user device and the communication device 14 are wirelessly connected. The user device can be a smartphone, tablet, laptop, or other electronic device capable of viewing a webpage interface.

[0021] In one embodiment of the present invention, when the area ratio is less than a preset value, the controller 13 determines the plant health as abnormal and sends a text message or email notification to the user device through the communication device 14. The user can react immediately after receiving the notification, such as providing supplementary light source or water. The preset value can be 70%, 50% or 30%. Those skilled in the art can adjust the size of the preset value according to the type of plant. The present invention is not limited to the above values.

[0022] The plant health assessment device 10 of one embodiment of the present invention further includes a basic light source 15, which is used as an illumination light source when taking assessment images, or as a light source for normal plant growth. The basic light source 15 can be a full-spectrum light source, and when it is used as an illumination light source during shooting, its intensity is one-tenth of that of the light source for plant growth.

[0023] Please refer to Figure 3, a flowchart of a plant health assessment method according to an embodiment of the present invention, and Figures 4A-4D, schematic diagrams of various images in a plant health assessment method according to an embodiment of the present invention. The plant health assessment method of this embodiment is applicable to the plant health assessment device 10. The plant health assessment method includes the following steps: S12: Ultraviolet light source 11 provides ultraviolet light to the plant leaf surface; S13: RGB camera 12 captures the plant leaf surface to obtain an assessment image (as shown in Figure 4A); S14: Controller 13 divides the assessment image into multiple pixels (as shown in Figure 4B), and separates each pixel into a red light component R, a green light component G, and a blue light component B; S15: Controller 13 calculates the ultraviolet normalization difference plant index (uNDVI) value of each pixel based on the separated components, where the uNDVI value is (RB) / (R+B); and S16: Controller 13 generates a uNDVI image based on the uNDVI value calculated for each pixel to assess plant health (as shown in Figure 4C).

[0024] More preferably, the ultraviolet light source 11 in one embodiment of the present invention can be a UVA light source with a wavelength of 320nm-400nm, and more preferably an ultraviolet light source 11 with a wavelength of 365nm.

[0025] More preferably, a plant health assessment method according to an embodiment of the present invention further includes step S17: the controller 13 calculates the area ratio of all pixels with uNDVI values ​​greater than or equal to a specific value to the plant leaf surface. The plant health is determined based on the area ratio, wherein the uNDVI value and the specific value are between -1 and 1. Furthermore, those skilled in the art can adjust the size of the specific value according to the plant species, and the present invention does not limit this specific value.

[0026] More preferably, step S17 further includes grayscale mapping processing. The controller 13 converts the uNDVI value to a grayscale value and the specific value to a specific grayscale value via grayscale mapping processing. The controller 13 also calculates the proportion of grayscale areas where the grayscale value is greater than or equal to the specific grayscale value to determine plant health. The grayscale value and the specific grayscale value are between grayscale 0 and 255. Furthermore, those skilled in the art can adjust the size of the specific grayscale value according to the plant species; this invention does not limit the specific grayscale value.

[0027] In one embodiment of the plant health assessment method of the present invention, when the area ratio of all pixels with uNDVI values ​​greater than or equal to a specific value to the plant leaf surface is greater than 90%, or when the grayscale area ratio of all pixels with grayscale values ​​greater than or equal to a specific grayscale value to the plant leaf surface is greater than 90%, the controller 13 determines the plant health to be good. When the area ratio or grayscale area ratio is between 70% and 90%, the controller 13 determines the plant health to be normal. When the area ratio or grayscale area ratio is less than 70%, the controller 13 determines the plant health to be abnormal. Those skilled in the art can adjust the standards used by the controller 13 to determine the plant health as good, normal, or abnormal according to the type of plant; the present invention is not limited to the above values.

[0028] More preferably, in a plant health assessment method according to an embodiment of the present invention, step S13 further includes: after obtaining the assessment image, the controller 13 first uses AI image recognition technology to determine the plant leaf surface in the assessment image and excludes the non-plant leaf surface parts in the assessment image, and then calculates the uNDVI value (i.e., steps S14 and S15) to make the calculation of the area ratio more accurate.

[0029] More preferably, the plant health assessment method of one embodiment of the present invention further includes step S18: the controller 13 performs pseudo-color rendering on the uNDVI image to display the plant health of the plant leaves (as shown in Figure 4D), and imports an AI model to determine the types of pests and diseases in pixels with uNDVI values ​​less than a specific value. For example, green can be used to mark pixels with uNDVI values ​​greater than or equal to a specific value, and blue can be used to mark pixels with uNDVI values ​​less than a specific value. Therefore, the user can intuitively see where the leaf surface in the assessment image is a defective leaf surface. In addition, the import of the AI ​​model can also enable the controller 13 to more accurately identify the area of ​​the defective leaf surface.

[0030] More preferably, the plant health assessment method of one embodiment of the present invention further includes step S19: providing a webpage interface showing the display area ratio of the user device through the communication device 14, and sending a prompt light corresponding to the plant health. When the controller 13 determines that the plant health is good, the prompt light is green; when the controller 13 determines that the plant health is normal, the prompt light is yellow; when the controller 13 determines that the plant health is abnormal, the prompt light is red.

[0031] More preferably, the plant health assessment method of one embodiment of the present invention further includes step S20: when the area ratio is less than a preset value, the controller 13 determines the plant health as abnormal and sends a text message or email notification to the user device through the communication device 14. By real-time monitoring and issuing notifications when abnormalities occur, the user can react to abnormal plant conditions at any time, wherein the preset value can be 70%, 50%, or 30%.

[0032] More preferably, the plant health assessment method of one embodiment of the present invention further includes step S11: the basic light source 15 provides light for illumination. When the light in the plant cultivation facility at night or indoors is extremely dim and the RGB camera 12 cannot capture a clear assessment image, the basic light source 15 can be used to provide illumination. The basic light source 15 can also be used as light for normal plant growth, and the light intensity when used for shooting illumination is one-tenth of the light intensity when used for plant growth.

[0033] In summary, the plant health assessment device and method of the present invention use an RGB camera with a controllable ultraviolet light source to capture images of plant leaves through the RGB camera. This allows for the reception of a large area of ​​leaf surface reflectance spectrum, analysis of chlorophyll activity and distribution to assess plant health, understanding of the overall plant health trend based on uNDVI values, and identification of defective leaves through uNDVI images. This can be used for real-time monitoring of plant growth, allowing for adjustments to environmental control conditions based on plant health status at any time.

[0034] Although the present invention has been disclosed above by way of embodiments, it is not intended to limit the present invention. Those skilled in the art to which the present invention pertains may make some modifications and refinements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims. [Simplified Explanation of the Diagram]

[0010] Figure 1 is a structural diagram of a plant health assessment device according to an embodiment of the present invention; Figure 2 is a schematic diagram of an application scenario of a plant health assessment device according to an embodiment of the present invention; Figure 3 is a flowchart of a plant health assessment method according to an embodiment of the present invention; and Figures 4A-4D are schematic diagrams of various images in a plant health assessment method according to an embodiment of the present invention.

Claims

1. A plant health assessment device, comprising: an ultraviolet light source for irradiating a plant leaf surface with ultraviolet light; an RGB camera for capturing images of the plant leaf surface and obtaining an assessment image; and a controller connected to the ultraviolet light source and the RGB camera for controlling the ultraviolet light source and the RGB camera, wherein... The controller divides the evaluation image into multiple pixels and separates each pixel into a red light component R, a green light component G, and a blue light component B. Based on the separated components, the controller calculates a UV normalization difference plant index (uNDVI) value for each pixel and generates a corresponding uNDVI image to determine the plant's health. The uNDVI value is (RB) / (R+B).

2. The plant health assessment device as described in claim 1, wherein the controller calculates the ratio of the area of ​​all pixels whose uNDVI value is greater than or equal to a specific value to one of the areas of the plant leaf surface to determine the plant health.

3. The plant health assessment device as described in claim 2, wherein the controller converts the uNDVI value to a grayscale value via a mapping grayscale process, converts the specific value to a specific grayscale value via the mapping grayscale process, and calculates the ratio of the area of ​​all pixels whose grayscale value is greater than or equal to the specific grayscale value to a grayscale area of ​​the plant leaf surface to determine the plant health.

4. The plant health assessment device as described in claim 1, wherein after acquiring the assessment image, the controller uses an AI image recognition technology to exclude parts of the assessment image that are not the leaves of the plant.

5. The plant health assessment device as described in claim 2 further includes a communication device coupled to the controller, the controller providing a web interface for a user device to display the area ratio via the communication device, and sending a prompt signal corresponding to the plant health.

6. A method for assessing plant health, applicable to a plant health assessment device, the device comprising an ultraviolet light source, an RGB camera, and a controller connected to the ultraviolet light source and the RGB camera, the method comprising: The ultraviolet light source provides ultraviolet light to the leaf surface of a plant; The RGB camera photographed the plant's leaves to obtain an evaluation image; The controller divides the evaluation image into multiple pixels and separates each pixel into a red light component R, a green light component G, and a blue light component B. Based on the separated components, the controller calculates a UV normalization difference plant index (uNDVI) value for each pixel and generates a corresponding uNDVI image to determine the plant's health. The uNDVI value is (RB) / (R+B).

7. The plant health assessment method as described in claim 6 further includes: The controller calculates the ratio of the area of ​​all pixels whose uNDVI value is greater than or equal to a specific value to that of one of the plant's leaf surfaces to determine the plant's health.

8. The plant health assessment method as described in claim 7, wherein the controller converts the uNDVI value to a grayscale value via a mapping grayscale process, converts the specific value to a specific grayscale value via the mapping grayscale process, and the controller calculates the ratio of the area of ​​all pixels whose grayscale value is greater than or equal to the specific grayscale value to a grayscale area of ​​the plant leaf surface to determine the plant health.

9. The plant health assessment method as described in claim 6 further includes: The controller uses AI image recognition technology to exclude parts of the evaluation image that are not the leaves of the plant.

10. The plant health assessment method as described in claim 7, wherein the plant health assessment device further includes a communication device coupled to the controller, the controller providing a web page interface for a user device to display the area ratio via the communication device, and sending a prompt signal corresponding to the plant health.

Images