CN102384767B - Nondestructive detection device and method for facility crop growth information - Google Patents

Abstract

The invention discloses a device and method for non-destructive detection of facility crop growth information, belonging to the technical field of facility crop monitoring. Including growth information sensing system, electronically controlled mechanical rocker and control computer; the control computer drives the electronically controlled mechanical rocker to locate the detection position, and controls the growth information sensing system, using multi-spectral imager, infrared temperature, radiation temperature, ambient temperature and humidity, and load sensors to obtain crop nitrogen, phosphorus, potassium, water reflection spectra, multispectral images, canopy temperature characteristics, multispectral morphological characteristics and fruit quality information of canopy, stems, plants, and fruits, as well as environmental Illumination, temperature and humidity information; optimize and compensate crop nutrition and water characteristics to obtain nutrition and water characteristic space; extract multi-spectral morphological characteristics of crops to obtain canopy area, stem diameter, fruit quality, plant height and other growth information , combined with nutrition, water and growth characteristics, the comprehensive acquisition and non-destructive detection of crop growth information is realized.

Description

Device and method for non-destructive detection of facility crop growth information

technical field

The invention relates to a non-destructive detection device and method for facility crop growth information, belonging to the technical field of facility crop monitoring. Specifically refers to the non-destructive detection of crop growth information using the visible light-near-infrared reflectance spectrum, multi-spectral image, canopy temperature, canopy illumination, ambient temperature and humidity, and load sensors of facility crops. The organic integration of multiple non-destructive detection technologies, combined with the monitoring of the light and temperature and humidity environment of crop growth, obtains the nitrogen, phosphorus, potassium, water, canopy area, stem thickness, plant height, fruit quality and other crop growth of facility crops synchronously in an all-round way Through the information provided by the device, water and fertilizer management and facility environment regulation can be carried out according to the growth status and dynamic needs of facility crops.

Background technique

The growth information of facility crops refers to the comprehensive information including the levels of main nutrient elements, water content, growth of stems, fruits and leaves of crops.

The total area of protected horticulture in my country ranks first in the world, and the proportion of large-scale multi-span greenhouses representing the modernization level of protected horticulture increases year by year. However, due to the lack of advanced and applicable facility crop fertilizer and water demand information and growth information detection system, it is impossible to conduct comprehensive and accurate detection and analysis of crop nutrition, water content and growth information, and it is impossible to perceive and reflect the real control requirements of crop growth, resulting in Crop yield potential is not fully exploited. In order to obtain high-yield and high-quality products and high economic returns, it is necessary to obtain a series of crop growth information based on the actual needs of the crops, and conduct a comprehensive evaluation of the growth information, and carry out water and fertilizer management and environmental regulation based on the evaluation results. Therefore, there is an urgent need to apply more comprehensive, systematic and scientific crop growth information detection devices to guide the production and operation of modern facilities to increase production, reduce regulation costs, reduce waste and non-point source pollution caused by excessive fertilization, and improve economic benefits.

The non-destructive detection of facility crop growth information mainly includes the detection of crop nitrogen, phosphorus and potassium nutrients, water and other nutrients and the detection of growth information such as canopy area, stem diameter, plant height, plant growth rate, fruit quality and growth rate. Due to the lack or excess of crop nutrition and water, the physiological characteristics of the surface and internal tissues of crop leaves will change, which will cause changes in the spectral reflection characteristics and image characteristics of crop leaves and canopies. At the same time, the water stress state of crops is significantly related to its canopy temperature. By analyzing the change law of the canopy-air temperature difference (difference between canopy temperature and ambient temperature) under different saturated water vapor pressures, the water information of plants can be detected non-destructively. The growth information such as canopy area, stem thickness, plant height, and fruit appearance of crops can also be obtained through image methods.

At present, there have been some related researches on crop nutrition and moisture detection. The lack and excess of crop nutrients will cause changes in the spectral reflectance characteristics of crops. Based on this principle, in terms of spectral detection, the invention patent application with application number 200510088935.0 discloses a portable non-destructive detection method and measuring instrument for plant nitrogen and water content, by detecting the spectra of plant leaves at four characteristic wavelengths The reflection intensity information is used to diagnose the nutrition of plants, and the nitrogen and water content information of plants is obtained by inversion of the vegetation index of four wavelengths. The utility model patent application with the application number of 200820078489.4 discloses a nitrogen reflectance index detector, which uses the spectral reflectance information of crop leaves at two specific wavelengths as the nitrogen reflectance index to infer the yield and quality of the crop. The invention patent application with application number 200510088935.0 discloses a non-destructive detection method for physiological indicators of plant leaves, which can detect chlorophyll, lutein, nitrogen and water by using spectral reflection information at 380-1100nm. In the invention patent application with the application number 200410048127.7, a method for predicting the nitrogen content of cucumber leaves based on the natural light reflection spectrum is disclosed. The reflection vegetation index of the leaves can be obtained from the spectral reflection intensity of the cucumber leaves at a specified wavelength, and then judged its nitrogen level. At present, the research method involved in the patent of spectral diagnosis of nutrient information such as crop nutrition and water mainly uses the spectral reflectance and combination information of plant leaves at certain wavelengths to detect the nutrition of crops, that is to say, through Analyze the spectral reflectance of a single leaf, and then infer the nutritional status of a single plant, and analyze the group nutritional level of plants in the area accordingly. However, only using the leaf information of a plant cannot fully characterize the nutritional status of the entire plant. Therefore, inferring the nutritional level of a crop from a single leaf will cause a large error. Therefore, nutritional diagnosis methods based on the canopy level can really meet the needs. In addition, because the spectral information collection adopts the point source sampling method, which has high requirements on the sampling points and is easily affected by background and environmental factors. Therefore, only using the reflectance spectral information of crops for nutrient and water detection errors is relatively large.

The visual image detection of crop nutrition is based on the changes in physical characteristics caused by crop nutrition deficiency, and uses image sensors to obtain crop color (gray scale), texture and other characteristics related to nutrition levels. In terms of visual image detection, the invention patent application number is 200710069116.0, which discloses a method for quickly and non-destructively measuring the nitrogen content of tea trees with multi-spectral imaging technology. The invention patent application with the application number 200510062298.X and the utility model patent application with the application number 200520134360.7 disclose a multi-spectral image diagnosis method and diagnosis system for rapeseed nitrogen nutrition. The above systems all use 3CCD multi-spectral camera system as the visual acquisition device. Under computer control, the 3CCD multi-spectral camera system collects multi-spectral image information of plant canopy, which can non-destructively diagnose the nitrogen nutrition status of plants. Although this type of system can diagnose the nitrogen nutrition status of plants through the analysis of the color and texture characteristics of the multi-spectral image of the plant canopy, due to the interaction between plant nutrition, especially the relationship between nitrogen and water There is an obvious positive interaction, and this type of system cannot detect the water stress information of the plant, so the detection of nitrogen will also be affected to a certain extent when the water information is not known. Moreover, the 3CCD multispectral camera system can only obtain multispectral image information of a specific wavelength, and it is difficult to accurately and effectively extract and screen the nutritional characteristics of crops. Therefore, only multispectral or hyperspectral with variable wavelengths and higher spectral resolution The imaging system can accurately obtain the nutrient information of crops and improve the accuracy of nutrient detection.

In terms of crop moisture detection using canopy temperature. The invention patent application with the application number 200710178192.5 and the utility model patent application with the application number 200720190401.3 disclose an online crop crown air temperature difference irrigation decision-making monitoring system, through a set of infrared canopy temperature sensors and brackets installed inside the high-speed pan/tilt The ambient temperature sensor and other monitoring devices set on the pole can realize the monitoring of the canopy temperature of the crops in the plot. However, since the water stress index index based on the canopy temperature difference can only indicate the trend of water stress, it cannot quantitatively evaluate the water content of the plant, and the water stress index is greatly affected by the environmental temperature and humidity, so it must be carried out using the environmental information collected synchronously. Correction in real time. Therefore, only using a single canopy temperature difference information to monitor plant moisture can only be used to judge trends. Therefore, it is necessary to introduce various features such as near-infrared reflection spectrum and image information, and to correct environmental factors in real time in order to improve plant moisture. Detection accuracy.

In terms of crop growth detection, the invention patent application with the application number 200610097576.X discloses an embedded agricultural plant growth state monitor and its working method, which can monitor the temperature and humidity of the crop growth environment, stem thickness, plant height, Soil viscosity and pH are detected. The system only judges crop growth by stem thickness and plant height, and lacks a dynamic crop growth evaluation model, so it is difficult to make a comprehensive and scientific evaluation of crop growth. The invention patent application with the application number 200410014648.0 discloses a device and method for crop growth monitoring and nutrient fertilization prescription generation. It is difficult to accurately analyze the nitrogen, phosphorus and potassium nutrition and water characteristics of crops because the camera can only obtain synthetic images of visible light. Although the nutrient composition detector can obtain the nutritional information of crops, its sampling and The detection method can cause damage to the crop.

At present, the non-destructive detection of crop growth information is mainly based on spectral and image technology. Spectral technology can easily obtain the relationship between nitrogen content, water content and spectral reflectance or its evolution; visible light or near-infrared visual image color (gray scale), texture, and morphological features can also represent crop nutrition levels to a certain extent , water stress, leaf area, stem and fruit leaves and other information, the crown-air temperature difference of crops is also significantly correlated with water stress. However, it is difficult to make a comprehensive, systematic and scientific judgment on the growth status of crops with only a single detection method of spectrum, image and canopy temperature to obtain isolated information such as nutrition or water or leaf area index, stem and fruit weight. Moreover, there is an interaction between nutrients and between nutrients and water. The detection process is greatly affected by environmental factors such as crop canopy structure, soil background spectrum, atmospheric window, temperature and humidity. Therefore, only spectral technology or visible light visual image , or near-infrared visual images, or a single detection technology such as the canopy temperature difference of a plant is not enough to accurately and comprehensively reflect the growth information of crop nutrition, water and growth.

In summary, there is an urgent need for facility crop growth information detection devices that can integrate multiple non-destructive testing technologies and comprehensively use multiple information to quickly extract, accurately analyze and scientifically evaluate growth information such as crop nutrition, water, and growth. , scientifically guide facility environment regulation and water and fertilizer management, improve the quality and output of agricultural products, reduce the cost of facility environment regulation, in order to improve economic benefits and realize the leapfrog development of facility horticulture in my country.

Contents of the invention

The purpose of the present invention is to provide a multi-sensing information fusion technology, which can make full use of various effective information such as visible light-near infrared reflection spectrum, multi-spectral image, canopy temperature, canopy illumination, environmental temperature and humidity of facility crops , a facility crop growth information rapid non-destructive detection system that comprehensively judges and scientifically evaluates the nutrition, water level and growth information of crops, providing a scientific basis for modern facility environment regulation and water and fertilizer management.

In order to achieve the above purpose, a device and method for non-destructive detection of facility crop growth information in the present invention adopts the following technical solutions:

A facility crop growth information non-destructive testing device, including three parts of an electronically controlled mechanical rocker, a growth information sensing system and a control computer, wherein the growth information sensing system and the control computer are installed on the electronically controlled mechanical rocker;

The electronically controlled mechanical rocker arm includes a tripod, a vertical lifting rod, a horizontal moving rod, an electronically controlled pan-tilt A and an electronically controlled pan-tilt B; the tripod is installed at the bottom of the electronically controlled mechanical rocker, and three universal wheels are installed at the bottom , There is a fixed internal thread hole in the center of the upper end of the tripod, and a shaft sleeve in the center of the lower end, which is socketed to install the vertical lifting rod; the vertical lifting rod is a screw structure, and its top is connected to the horizontal moving rod through a cross connector; the horizontal moving rod is a screw structure, Located at the top of the vertical lifting rod and the top of the electronically controlled mechanical rocker arm; the electronically controlled pan/tilt A is installed on the lead screw of the horizontal moving rod through the internal thread slider, and the electronically controlled pan/tilt B is installed on the horizontal moving rod through the internally threaded slider on the screw rod of the vertical lift rod.

The growth information sensing system includes a multi-sensor unit, a data acquisition card and a light source. The multi-sensor unit is installed on the electronically controlled pan-tilt A and the electronically controlled pan-tilt B of the electronically controlled mechanical rocker arm; the light source is installed on the electronically controlled pan-tilt A. , located directly below the multi-sensor unit; the data acquisition card is connected with the multi-sensor unit, and installed above the tripod of the electromechanical rocker arm.

The control computer is installed above the tripod of the electronically controlled mechanical rocker, and is connected with the data acquisition card, the motion control card, the multispectral imager A, and the multispectral imager B through a USB bus.

The multi-sensor unit described in the present invention includes a multispectral imager A, a multispectral imager B, an infrared temperature detector A, an infrared temperature detector B, an infrared temperature detector C, an irradiance sensor, a temperature and humidity sensor, a load Sensor, shading cover and ruler; Wherein multi-spectral imager A, infrared temperature detector A, infrared temperature detector B, irradiance sensor, temperature and humidity sensor, shading cover are installed on the electronically controlled cloud of described electronically controlled mechanical rocker On the platform A, it is fixed under the electric control platform A, and is in a overlooking position; the multi-spectral imager B and the infrared temperature detector C are installed on the electronic control platform B of the electronically controlled mechanical rocker, fixed on the electronically controlled The left side of the cloud platform B is in a side-view position; the load sensor is located below the fruit of the test sample, and is vertically fixed in the greenhouse soil tank through the support rod; the scale is fixed beside the test sample vertically to the ground and parallel to the test sample .

A method for non-destructive detection of facility crop growth information in the present invention is carried out according to the following steps:

(1) Use multispectral imager A and multispectral imager B to collect visible light-near-infrared reflectance spectra and multispectral images of the top-view and side-view fields of test samples, and upload them to the control computer through the USB bus, based on which it can be judged Detect nitrogen, phosphorus, and potassium nutrient levels, canopy area, and fruit shape information of samples;

(2) Use infrared temperature detector A, infrared temperature detector B, and infrared temperature detector C to collect the canopy temperature information of the test sample; use the load sensor to collect the fruit quality information of the test sample; use the irradiance sensor and temperature and humidity The sensor collects and detects the light, temperature and humidity information of the growth environment of the sample; the above information is input into the data acquisition card for digital conversion, and then uploaded to the control computer through the USB bus;

(4) Analyze and process the collected visible light-near-infrared reflectance spectrum and multispectral images, and control the computer to extract the spectral characteristic wavelengths of nitrogen, phosphorus and potassium of the test samples and the color, texture, gray average and fusion characteristics of multispectral images , use the synchronously acquired light intensity information to perform feature compensation, and then optimize the extracted spectral characteristic wavelength of NPK, the color, texture, gray value and fusion features of the multi-spectral image, and construct the NPK spectrum and image combination features space;

(5) Using the collected visible light-near-infrared reflectance spectrum and the canopy temperature information of the detected samples, control the computer to extract the spectral characteristics and canopy temperature characteristics of the detected sample moisture, and combine the environmental temperature and humidity information to obtain the canopy-air temperature difference and saturated water vapor pressure; use the synchronously acquired environmental light information to perform feature compensation, and construct the combined feature space of water spectrum and canopy temperature through feature optimization;

(6) Using the collected multi-spectral images and fruit quality information, combined with the reference scale, control the computer to extract the canopy area, stem diameter, plant height, fruit shape and quality data of the test samples; and obtain the canopy Area expansion rate, plant growth rate and fruit growth rate;

(7) Use the obtained crop nitrogen, phosphorus, potassium nutrition, water and growth information, and use the control computer to continuously monitor and record, as the detection data of the growth information of the detection sample.

The analysis and processing method of the reflectance spectrum adopted in step (4) includes: first filtering, and then performing stepwise regression and principal component analysis; the analysis and processing method of the multispectral image includes: first enhancing the multispectral image and performing pixel-level Image fusion, and then segment the background through super green features and two-dimensional histograms, and finally perform color (grayscale) mean calculation, texture analysis, and fusion feature analysis.

The effects of the present invention are (1) the present invention simultaneously uses the visible light-near-infrared reflectance spectrum and multi-spectral images of crops, canopy temperature, canopy illumination, ambient temperature and humidity, and load sensors to perform non-destructive detection of crop growth information, and can detect crop growth information through spectral , image, infrared temperature, radiation intensity and other non-destructive detection technologies, combined with the environment of crop growth and temperature and humidity monitoring, all-round synchronous acquisition of comprehensive information on crop growth, not only the amount of information acquired is larger and richer, but also able to A more comprehensive and accurate grasp of the growth status of crops has not been covered in previous documents; (2) Traditional nutrient detection methods usually only detect nitrogen nutrition and water, and the growth status information is only based on leaf area, stem diameter and Due to the interaction and antagonism between crop nutrients and the interaction between nutrients and water, and the uncertainty of crop growth and its causes, it is only based on the nitrogen, water or plant height of crops. , The detection of stem diameter information is not enough to fully reflect the growth status of crops. Compared with the traditional single detection method and detection object, the present invention obtains nutrient information such as nitrogen, phosphorus, potassium and water of crops, and growth information such as canopy area, stem diameter, plant height, fruit quality, plant and fruit growth rate, etc., and combines Environmental information detection can carry out continuous and comprehensive monitoring of the growth status of facility crops, which has not been involved in previous documents; (3) The present invention integrates the visible light-near-infrared reflection spectrum and multi-spectral image information of crops, To comprehensively judge the nitrogen, phosphorus and potassium nutrition level of crops; judge the water stress state of crops through the information fusion of near-infrared spectrum and canopy infrared temperature, and judge the growth of crops by extracting morphological features from visual images of crops, and judge the growth of crops through environmental light Compensate the measurement error with the temperature and humidity information to obtain the growth information of the crops comprehensively and accurately. Comprehensive growth information improves efficiency and reduces labor intensity. the

Description of drawings

Fig. 1 is a structural schematic diagram of a non-destructive detection device for facility crop growth information according to the present invention;

1-ruler 2-shade 3-halogen light source A

4-Irradiance sensor 5-Temperature and humidity sensor 6-Infrared temperature detector A

7-Multi-spectral imager A 8-Infrared temperature detector B 9-Halogen light source B

10-Electrically controlled pan/tilt A 11-Multi-spectral camera B 12-Infrared temperature detector C

13-Load sensor 14-Detection sample 15-Horizontal moving rod

16-Vertical Lifting Rod 17-Electrically Controlled Pan Tilt B 18-Electrically Controlled Mechanical Rocker Arm

19-Tripod 20-Data Acquisition Card 21-Motion Control Card

22 - Control computer 23 - Greenhouse soil tank.

Detailed ways

The present invention will be described in detail below in conjunction with the accompanying drawings.

Referring to the accompanying drawing 1, a facility crop growth information non-destructive testing device of the present invention includes three parts: an electronically controlled mechanical rocker 18 , a growth information sensing system and a control computer 22 . The growth information sensing system and the control computer 22 are installed on the electronically controlled mechanical rocker arm 18, and the front, rear, left and right, and up and down positions are adjusted by controlling the mechanical rocker arm 18, so that the growth information sensing system reaches the detection position.

The electronically controlled mechanical rocker arm 18 includes a tripod 19, a vertical lifting rod 16, a horizontal moving rod 15, an electronically controlled platform A10 and an electronically controlled platform B17. Described tripod 19 is installed on the bottom of electric control mechanical rocker 18, and its bottom end is equipped with three universal wheels, and effect is to stabilize electric control mechanical rocker 18, and can make electric control mechanical rocker 18 move back and forth, left and right in working direction ; The center of the upper end of the tripod has a fixed internal threaded hole, and the center of the lower end has a shaft sleeve, and the vertical lifting rod 16 is socketed and installed; the vertical lifting rod 16 is a screw structure, and its top is connected to the horizontal moving rod 15 by a cross connector; The screw structure is located at the top of the vertical elevating rod 16 and at the top of the electronically controlled mechanical rocker arm 18; the electronically controlled pan/tilt A10 is installed on the leading screw of the horizontal moving rod 15 through an internal thread slider, and the electronically controlled pan/tilt B17 passes through The internal thread slide block is installed on the screw mandrel of vertical elevating rod 16. The electronically controlled mechanical rocker arm 18 can rotate and drive the internally threaded slider through the screw of the vertical lifting rod 16 to drive the electronically controlled pan-tilt B17 to adjust the position in the vertical direction within the height range of 500-3000 mm; the electrically controlled mechanical rocker arm 18 can be Through the rotation of the lead screw of the horizontal moving rod 15, the inner threaded slider is driven to drive the electric control pan-tilt A10 to adjust the horizontal position within the range of 200-2000 mm. Through the front-back and left-right displacement movements of the electronically controlled mechanical rocker arm 18, combined with the horizontal and vertical position adjustments of the electronically controlled pan/tilt A10 and the electronically controlled pan/tilt B17, the growth information sensor mounted on the electronically controlled mechanical rocker 18 can be driven. The system reaches the desired detection position.

The growth information sensing system includes a multi-sensor unit, a data acquisition card 20 and a light source. The multi-sensor unit is installed on the electronically controlled pan-tilt A10 and the electronically controlled pan-tilt B17 of the electronically controlled mechanical rocker arm 18. Just below the multi-sensor unit part installed on the top; the data acquisition card 20 is connected with the multi-sensor unit, and is installed above the tripod 19 of the electromechanical rocker arm 18. The multi-sensor unit includes a multispectral imager A7, a multispectral imager B11, an infrared temperature detector A6, an infrared temperature detector B8, an infrared temperature detector C12, an irradiance sensor 4, a temperature and humidity sensor 5, and a load sensor 13 , Shade 2 and Ruler 1. Wherein the multi-spectral imager A7, the infrared temperature detector A6, the infrared temperature detector B8, and the hood 2 are installed on the electronically controlled pan-tilt A10 of the electronically controlled mechanical rocker arm, and are fixed under the electronically controlled pan-tilt A10. Looking down, the multi-spectral imager A7 is installed at the center of the electronic control platform A, and the infrared temperature detector A6 and infrared temperature detector B8 are installed at a quarter of the center line of the rectangle below the electronic control platform A10 The symmetrical positions are located on the left and right sides of the multi-spectral imager A7 respectively; the irradiance sensor 4 and the temperature and humidity sensor 5 are fixed at the symmetrical positions of one-third of the two ends of the center line of the rectangle above the electronically controlled pan-tilt A10 . The multi-spectral imager B11 and the infrared temperature detector C12 are installed on the electronically controlled pan-tilt B17 of the electronically controlled mechanical rocker arm, and fixed at the symmetrical positions of one-third of the two ends of the rectangular center line on the left side of the electronically controlled pan-tilt B17, In side looking position, multispectral imager B11 is located above infrared temperature detector C12. The load sensor is located under the fruit of the detection sample 14, and is installed on the support rod, and the support rod is fixed on the greenhouse soil tank 23, perpendicular to the greenhouse soil tank 23; the scale 1 is fixed at 20 cm on the left side of the detection sample 14, vertically It is parallel to the greenhouse soil tank 23 and the detection sample 14 . Among the sensors mentioned above, the power input ends of infrared temperature detector A6, infrared temperature detector B8, infrared temperature detector C12, irradiance sensor 4, temperature and humidity sensor 5, and load sensor 13 are connected in parallel, all using DC24V Power supply, its output is all ~ 5V signal, and the signal output end is all connected with the signal input end of data acquisition card 20, uploads control computer 22 through the USB bus line of data acquisition card 20 after carrying out digital conversion; The spectral imager B11 is connected to the control computer 22 through a USB bus, and uses the USB bus power supply and data transmission of the control computer to upload the collected visible light-near infrared reflection spectrum and multi-spectral image information to the control computer 22 .

The multi-spectral imager A7 in the overlooking position can simultaneously acquire the visible light-near-infrared reflection spectrum and multi-spectral image information of the detection sample 14 at five different characteristic wavelengths in the wavelength range of 300~1100nm, based on which the crop quality can be judged. Nitrogen, phosphorus, potassium nutrient levels and canopy area, fruit shape information. The multi-spectral imager B11 in the side-view position is the same model as the multi-spectral imager A7. Combined with scale 1, it can obtain visible light-near-infrared reflectance spectra and multi-spectral images of crop stalks, plants, and test samples at different angles of view. Information, based on which the stem diameter, plant height information and NPK nutritional information of crops can be obtained. Infrared temperature detector A6, infrared temperature detector B8, and infrared temperature detector C12 are of the same model, with a measuring range of -40 to 80°C. The acquired environmental temperature and humidity information can be used to obtain the canopy-air temperature difference and saturated water vapor pressure, based on which the water stress state of the test sample can be judged. 0-100% RH; the irradiance sensor 4 is used to obtain the real-time light intensity, which can compensate the impact of ambient light intensity changes on the measurement accuracy, and its measurement range is 0-100Klux; the measurement range of the load sensor 13 is 0-1000g, the weight is fixed in the soil tank of the greenhouse by the support rod, the mass of the fruit of the crop sample 14 is measured by lifting it, and then the growth rate of the fruit is calculated according to the continuous measurement data.

Described data acquisition card 20 is 16 USB data acquisition cards, is installed on the tripod 19 tops of electromechanical rocking arm 18, and its signal input terminal and infrared temperature detector A6, infrared temperature detector B8, infrared temperature detector C12, The irradiance sensor 4, the temperature and humidity sensor 5, and the load sensor 13 signal output terminals are connected to each other, and the input signal is A/D converted, and its output terminal is connected to the control computer 22 through the USB bus of the data acquisition card 20, and the digital signal is uploaded to The control computer 22 performs analysis and processing.

The light source system includes a light source and a hood 2, which are used to provide the multispectral imager A7 and the multispectral imager B11 with a visible light-near-infrared light source in the 250-3000nm band to obtain clear multispectral images and stable reflection spectra data. The light source is composed of halogen light source A3 and halogen light source B9, which are respectively installed on the inside of the left and right ends of the rectangular center line of the slider connected by the electric control pan-tilt A10 and the horizontal moving rod 15; the hood 2 is installed on the horizontal moving rod 15 The outer sides of the rectangular centerline of the slider are used to shield the external light source, and enhance the radiation intensity and uniformity of the light source through the curved reflective surface of diffuse reflection, so as to shield external interference and improve the reflection spectrum and hyperspectral image Data accuracy and stability.

The main functions of the control computer 22 are to realize signal acquisition control, electronically controlled mechanical rocker arm motion control and data analysis. The control computer 22 is installed above the tripod 19 of the electronically controlled mechanical rocker arm 18, and is connected with the data acquisition card 20, the motion control card 21, the multispectral imager A7 and the multispectral imager B11 through a USB data cable, and controls the multispectral imaging Instrument A7 and multispectral imager B11 collect visible light-near-infrared reflection spectrum and multispectral image information of test samples; collect infrared temperature detector A6, infrared temperature detector B8, infrared temperature detector C12, radiation by controlling data acquisition card 20 Illuminance sensor 4, temperature and humidity sensor 5, load sensor 13 signal; control the motion of electronically controlled mechanical rocker 18 through motion control card 21. And display, analyze and process the obtained crop growth information data.

Steps for implementing a method for non-destructive detection of facility crop growth information:

(1) The control computer 22 controls the electronically controlled mechanical rocker arm 18 to adjust the front and rear positions on the inter-row operation channel between the facilities and crops. After reaching the detection sample position, it drives the screw of the vertical lifting rod 16 to rotate and drives the internal thread slider And the electronically controlled pan-tilt B17 makes the sensor installed on it go up and down to adjust the height position, and then adjust the horizontal position of the sensor by controlling the electronically controlled pan-tilt A10 on the horizontal moving rod 15 to reach the detection position directly above the crop;

(2) Start the halogen light sources 3 and 9, and start the multi-spectral imager A7 on the electronically controlled pan-tilt A10 directly above the test sample 14 and the multi-spectral imaging at the lateral horizontal position on the electronically controlled pan-tilt B17 in the middle of the crop respectively Instrument B11, collecting the visible light-near-infrared reflection spectrum and multispectral image of the top view and side view field of view of the detection sample 14, and uploading the control computer 22 through the USB bus;

(3) Start the infrared temperature detector A6, the infrared temperature detector B8 on the electronically controlled pan-tilt A10 directly above the test sample 14, and the infrared temperature detector C12 at the lateral horizontal position on the electronically controlled pan-tilt B17 in the middle of the test sample 14 , obtain the canopy temperature information of the crop; start the load sensor 13 located under the fruit of the detection sample 14 to obtain the fruit weight information; start the irradiance sensor 4 and the temperature and humidity sensor 5 located directly above the detection sample 14 synchronously, and collect the detection sample 14 The real-time light and ambient temperature and humidity information of the growth environment, and the differential voltage signal of the above-mentioned sensors is input to the data acquisition card 20 for A/D conversion and then uploaded to the control computer 22 through the USB bus;

(4) Use the collected visible light-near-infrared reflectance spectrum and multispectral image for processing. The analysis and processing methods of the reflectance spectrum include: first filter, then stepwise regression and principal component analysis; the analysis and processing method of multispectral image Including: firstly enhance the multi-spectral image and perform pixel-level image fusion, then segment the background by ultra-green features and two-dimensional histogram, and finally perform color (grayscale) mean value calculation, texture analysis and fusion feature analysis. Control the computer to extract the spectral characteristic wavelength of the nitrogen, phosphorus and potassium of the detection sample 14 and the color, texture, gray average and fusion features of the multi-spectral image, and use the synchronously acquired light intensity information to perform feature compensation, and then the extracted nitrogen, phosphorus, and potassium Optimizing the characteristic wavelength of the spectrum, the color, texture, gray value and fusion features of the multi-spectral image, and constructing the characteristic space of nitrogen, phosphorus and potassium spectrum and image combination;

(5) Using the collected visible light-near-infrared reflection spectrum and canopy temperature information, the control computer 22 extracts the reflection spectrum characteristics and canopy temperature characteristics of the water in the detection sample 14, and combines the environmental temperature and humidity information to obtain the canopy-air temperature difference and saturation Water vapor pressure; use the synchronously acquired environmental light information to perform feature compensation, and construct the combined feature space of water spectrum and canopy temperature through feature optimization;

(6) Using the collected multi-spectral image and fruit quality information, combined with the reference scale 1 reference target, the control computer 22 extracts the canopy area, stem diameter, plant height, fruit shape and quality data of the test sample 14; and according to the continuous observation data , get the canopy area expansion rate, plant growth rate and fruit growth rate;

(7) Using the acquired NPK nutrition, moisture and growth information of the test sample 14, the control computer 22 is used for continuous monitoring and recording as the test data of the growth information of the test sample 14.

Claims (3)

1. A facility crop growth information nondestructive testing device, characterized in that it includes three parts: an electronically controlled mechanical rocker, a growth information sensing system and a control computer, wherein the growth information sensing system and the control computer are installed on the electronically controlled mechanical rocker superior; The electronically controlled mechanical rocker arm includes a tripod, a vertical lifting rod, a horizontal moving rod, an electronically controlled pan-tilt A and an electronically controlled pan-tilt B; the tripod is installed at the bottom of the electronically controlled mechanical rocker, and three universal wheels are installed at the bottom , There is a fixed internal thread hole in the center of the upper end of the tripod, and a shaft sleeve in the center of the lower end, which is socketed to install the vertical lifting rod; the vertical lifting rod is a screw structure, and its top is connected to the horizontal moving rod through a cross connector; the horizontal moving rod is a screw structure, Located at the top of the vertical lifting rod and the top of the electronically controlled mechanical rocker arm; the electronically controlled pan/tilt A is installed on the lead screw of the horizontal moving rod through the internal thread slider, and the electronically controlled pan/tilt B is installed on the horizontal moving rod through the internally threaded slider On the screw rod of the vertical lifting rod; The growth information sensing system includes a multi-sensor unit, a data acquisition card and a light source. The multi-sensor unit is installed on the electronically controlled pan-tilt A and the electronically controlled pan-tilt B of the electronically controlled mechanical rocker arm; the light source is installed on the electronically controlled pan-tilt A. , located directly below the multi-sensor unit; the data acquisition card is connected to the multi-sensor unit, and installed above the tripod of the electromechanical rocker arm; The control computer is installed above the tripod of the electronically controlled mechanical rocker, and is connected with the data acquisition card, the motion control card, the multispectral imager A, and the multispectral imager B through a USB bus; The multi-sensor unit includes multispectral imager A, multispectral imager B, infrared temperature detector A, infrared temperature detector B, infrared temperature detector C, irradiance sensor, temperature and humidity sensor, load sensor, shading Cover and ruler; Wherein multispectral imager A, infrared temperature detector A, infrared temperature detector B, irradiance sensor, temperature and humidity sensor, shading cover are installed on the electric control cloud platform A of described electric control mechanical rocker arm , fixed on the bottom of the electronic control platform A, in the overlooking position; the multispectral imager B and infrared temperature detector C are installed on the electronic control platform B of the electronically controlled mechanical rocker, fixed on the electronic control platform B The left side is in a side view position; the load sensor is located under the test sample fruit, and is vertically fixed in the greenhouse soil tank through the support rod; the scale is fixed beside the test sample vertically to the ground and parallel to the test sample. 2. A non-destructive detection method for facility crop growth information, characterized in that it is carried out according to the following steps: (1) Use multispectral imager A and multispectral imager B to collect visible light-near-infrared reflectance spectra and multispectral images of the top-view and side-view fields of test samples, and upload them to the control computer through the USB bus, based on which it can be judged Detect nitrogen, phosphorus, and potassium nutrient levels, canopy area, and fruit shape information of samples; (2) Use infrared temperature detector A, infrared temperature detector B, and infrared temperature detector C to collect the canopy temperature information of the test sample; use the load sensor to collect the fruit quality information of the test sample; use the irradiance sensor and temperature and humidity The sensor collects and detects the light, temperature and humidity information of the growth environment of the sample; the above information is input into the data acquisition card for digital conversion, and then uploaded to the control computer through the USB bus; (4) Analyze and process the collected visible light-near-infrared reflectance spectrum and multispectral images, and control the computer to extract the spectral characteristic wavelengths of nitrogen, phosphorus and potassium of the test samples and the color, texture, gray average and fusion characteristics of multispectral images , use the synchronously acquired light intensity information to perform feature compensation, and then optimize the extracted spectral characteristic wavelength of NPK, the color, texture, gray value and fusion features of the multi-spectral image, and construct the NPK spectrum and image combination features space; (5) Using the collected visible light-near-infrared reflectance spectrum and the canopy temperature information of the detected samples, control the computer to extract the spectral characteristics and canopy temperature characteristics of the detected sample moisture, and combine the environmental temperature and humidity information to obtain the canopy-air temperature difference and saturated water vapor pressure; use the synchronously acquired environmental light information to perform feature compensation, and construct the combined feature space of water spectrum and canopy temperature through feature optimization; (6) Using the collected multi-spectral images and fruit quality information, combined with the reference scale, control the computer to extract the canopy area, stem diameter, plant height, fruit shape and quality data of the test samples; and obtain the canopy Area expansion rate, plant growth rate and fruit growth rate; (7) Use the obtained crop nitrogen, phosphorus, potassium nutrition, water and growth information, and use the control computer to continuously monitor and record, as the detection data of the growth information of the detection sample. 3. A method for non-destructive detection of facility crop growth information according to claim 2, characterized in that the analysis and processing of the reflectance spectrum adopted in step (4) includes: first filtering, followed by stepwise regression and principal component analysis; The analysis and processing method of the multispectral image includes: firstly, enhancing the multispectral image and performing pixel-level image fusion, then segmenting the background by ultra-green features and two-dimensional histogram, and finally performing color gray mean calculation, texture analysis and fusion feature analysis.