Study On The Method For Measuring Background-independent Vegetation Index Based On Narrow-banded Images

With the continuous development and progress of scientific and technological level,fine agriculture is moving towards all aspects of agricultural production.Due to the qualitative and quantitative relationship with the growth status of crops,the vegetation index is used as an important detection index for various intelligent detection Among the equipment,it has played a positive role in guiding the development of precision agriculture.In the field application,the vegetation index detection results of the existing commercial equipment are susceptible to background interference,especially in the early stage of crop growth,due to the low vegetation coverage,the soil-based background interference on the reflected spectrum signal of the canopy Has a very big impact.In order to effectively eliminate the influence of background interference on vegetation index measurement results,this paper designs and builds a set of anti-background interference plant canopy vegetation index measurement system,and proposes a vegetation index measurement method to remove background interference under field conditions.Using the characteristic that the proportion of radiant energy in each band of sunlight remains basically stable,using sunlight as the detection light source,the hardware and software systems are designed,and the image processing and stereo matching are applied to the algorithm.Anti-background interference canopy vegetation index measurement,and within the effective detection range of the design,the measurement results have high stability when measuring altitude,solar light intensity and solar altitude angle changes.The main research contents and research results of this article are as follows:(1)Analysis of research status at home and abroad:introduced the research status of crop canopy reflectance spectrometer at home and abroad,analyzed the advantages and disadvantages of the spectrometer developed at home and abroad,introduced the research status of image processing technology in crop nutrition.On this basis,determine the research goals of this article;(2)Measuring device software and hardware design and equipment calibration:According to functional requirements,a vegetation index measurement system constructed by three cameras with main optical axes parallel to each other and having the same performance parameters was developed.A filter was added to the camera lens to convert it to a narrow-band camera,and one of the near-infrared narrow bands was used.The gray value of the narrowband image obtained by the camera and a red light narrowband is used to calculate the vegetation index of the crop.The high contrast characteristics of the near infrared background image and the canopy image and the parallax information obtained by the two near infrared narrowband cameras of the same band are used to achieve background interference.Removal;for this measurement system,the corresponding image acquisition and processing software was developed,the internal and external parameter matrix of the camera was measured by a black and white checkerboard calibration plate,and the imaging geometric model was established;different heights and different heights were measured using standard whiteboards and grayscale plates The response test of the narrow-band camera under the sunlight intensity determined the measurement accuracy of the system using the image gray value to obtain the reflectance.The measurement results show that the gray plate with a nominal reflectance of 75%has an average reflectance of 75.47%in the 770nm band and a coefficient of variation of 1.5%;the average reflectance in the 660nm band is 75.56%and the coefficient of variation is 1.06%;nominal The gray plate with a 10%reflectance has an average reflectance of 9.33%and a coefficient of variation of 8.36%in the 770nm band;an average reflectance of 11.1%and a coefficient of variation of 5.01%in the 660nm band.It shows that the reflectance value obtained by the measuring device has a high consistency with the nominal reflectance value,and has a high measurement accuracy.Since the vegetation index is constructed based on the reflectivity,the method of obtaining the object gray information using the measuring device It can be used to calculate the vegetation index.(3)Calculation method of vegetation index without white board correction:Since the proportion of radiant energy in each band of sunlight is basically constant,a method of calculating the vegetation index using the standard whiteboard reflection response ratio ?of different detection bands is proposed.This method does not require whiteboard correction in actual measurement.The measurement height has nothing to do with the lighting environment.In different months(a total of 12 days are randomly selected in March,July,and December),different weather conditions(overcast,sunny,and cloudy),and different measurement time periods(from 9:00 to 17:00 in the morning,set 7 times Section)and different measurement heights(60cm,70cm,80cm,90cm and 100cm five measurement heights)were tested on the standard whiteboard reflection response ratio ? of the two detection bands of the measurement system.During the test,the change of the solar light intensity The range is 10000LUX?120000LUX.A single factor analysis of variance was performed on the measurement results obtained at different measurement time periods and measurement heights.The analysis results showed that there is no significant difference in the standard whiteboard reflection response ratio ?on the whole,which shows that the whiteboard reflection response ratio ? is basically not exposed to external light.Influence of condition and measurement height.For the test system constructed in this paper,the whiteboard reflection response ratio ? obtained during the test is ideal under the conditions of measuring height 70cm?100cm,measuring time 9:00?15:00,and solar light intensity ranging from 60000Lux to 110000Lux The test results,the mean value of ? is 1.0069,and the coefficient of variation is 7.62%..(4)Narrow-band image segmentation method:Using the high-contrast characteristics of the vegetation canopy and background of the narrow-band image of near-infrared light,a stable normalized threshold segmentation method was designed.For the crop canopy at the same measurement point and the same measurement height,when the solar light intensity changes from 30000LUX to 70000LUX,the normalized segmentation threshold is basically constant at about 0.35,and the coefficient of variation is 3.45%,indicating that this segmentation method can achieve near infrared Narrow-band image adaptive segmentation of the light environment;using near infrared narrow-band image segmentation results and spatial geometric relationships to achieve red light image canopy region segmentation through coordinate translation;evaluation of translation transformation results,the results indicate that red light narrow-band image The average correct rate of middle canopy segmentation is 88.53%,and the coefficient of variation is 2.13%,indicating that the canopy region of the red light image obtained by segmentation can effectively represent the overall situation of the canopy,and can be used to construct the vegetation index of the canopy.Taking the cultivated wheat canopy as the test object,nine measurement points were selected,and NDVI(Normalized Difference Vegetation Index)was conducted at different measurement heights(six measurement heights of 90cm,100cm,110cm,120cm,130cm and 140cm)located in the crop canopy.The results show that the NDVI measurement values at different altitudes and different solar intensities are basically stable,with a maximum coefficient of variation of 6.69%,indicating that the measurement method proposed in this paper is less affected by the measurement altitude and solar intensity.In this paper,based on a self-designed measuring device,a spectral vegetation index calculation method without whiteboard correction is constructed;a light intensity adaptive segmentation method for narrowband images is proposed,and the crop canopy and background in red light narrowband images are segmented using geometric transformation;The calculation method of using the gray-scale value of the narrow-band image as the reflection information of the measured object is determined;relevant experimental tests are carried out,and the results show that within the measurable range of the solar light intensity,the crop canopy vegetation index that can resist background interference under field conditions can be completed measuring.