Plant seedling quality assessment method and device based on remote sensing images

Abstract

The present invention provides a plant seedling quality assessment method and device based on remote sensing images, the method comprising: obtaining a remote sensing image of a target plot, inputting the remote sensing image into a trained target detection model, and obtaining a plant seedling recognition result; determining a seedling emergence area in the remote sensing image based on the plant seedling recognition result, the seedling emergence area being an area in the remote sensing image including seedlings of target plants in the target plot; performing Voronoi segmentation on the seedling emergence area in the remote sensing image based on the plant seedling recognition result to obtain a plurality of sub-areas; classifying the sub-areas based on the areas of the sub-areas to obtain a first sub-area, a second sub-area, and a third sub-area, and determining a seedling quality assessment result of a target plant in the target plot based on the areas of the sub-areas. The seedling quality assessment result includes a seedling deficiency rate and a seedling recovery rate, seedling emergence uniformity, and seedling vigor of a target crop in the target plot. The present invention can realize an automated assessment of seedling quality.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method and device for evaluating plant seedling emergence quality based on remote sensing images. Background Art

[0002] Management during the seeding and emergence phase of crops is crucial, as the quality of seedling emergence determines subsequent yield levels. However, poor seedling emergence quality can be caused by issues with seed emergence rates, machine or human operation, soil moisture and fertility, and underground pests, severely impacting final field yields and subsequent precise management. Accurate and timely understanding of seedling emergence across the entire field is a pressing need for both growers engaged in field crop production and breeders cultivating seed plots. Accurate, timely, and comprehensive assessment of seedling emergence quality provides effective information support for timely seedling replacement and subsequent management.

[0003] Traditional crop seedling quality assessment mainly relies on human inspections and experience-based judgments, which is time-consuming, labor-intensive, and unable to cope with a relatively large monitoring range. Summary of the Invention

[0004] The present invention provides a method and device for assessing the quality of plant seedlings at the seedling stage based on remote sensing images, which is used to overcome the defect in the prior art that manual seedling quality assessment cannot cope with a relatively large monitoring range, and realizes automated assessment of seedling quality.

[0005] The present invention provides a method for evaluating plant seedling emergence quality based on remote sensing images, comprising:

[0006] Obtaining a remote sensing image of a target plot, inputting the remote sensing image into a trained target detection model, and obtaining a plant seedling recognition result output by the target detection model, wherein the plant seedling recognition result includes coordinates of target plant seedlings in the target plot in the remote sensing image;

[0007] determining a seedling emergence area in the remote sensing image based on the plant seedling identification result, wherein the seedling emergence area is an area in the remote sensing image that includes the seedlings of the target plant in the target plot;

[0008] Performing Voronoi segmentation on the seedling emergence area in the remote sensing image based on the plant seedling identification result to obtain a plurality of sub-areas;

[0009] classifying the sub-regions based on their areas to obtain a first sub-region, a second sub-region, and a third sub-region, wherein the area of ​​the first sub-region is smaller than that of the second sub-region, and the area of ​​the second sub-region is smaller than that of the third sub-region;

[0010] The seedling emergence quality assessment result of the target plant in the target plot is determined based on the areas of the first sub-area, the second sub-area and the third sub-area. The seedling emergence quality assessment result reflects the seedling emergence quality of the target plant in the target plot. The seedling emergence quality assessment result includes the seedling deficiency rate and re-seedling rate, seedling emergence uniformity and seedling vitality of the target crop in the target plot.

[0011] According to the plant seedling emergence quality assessment method based on remote sensing images provided by the present invention, determining the seedling emergence area in the remote sensing image based on the plant seedling identification result includes:

[0012] Determining an initial plot in the remote sensing image based on the plant seedling identification result, the initial plot being a minimum convex polygonal plot including all the seedlings of the target plant in the target plot;

[0013] Expanding the vertex coordinates of the initial plot to obtain the vertices of the seedling emergence area;

[0014] Virtual points are inserted between adjacent vertices of the seedling emergence area to obtain a virtual boundary of the seedling emergence area.

[0015] According to the plant seedling emergence quality assessment method based on remote sensing images provided by the present invention, Voronoi segmentation is performed on the emergence area in the remote sensing image based on the plant seedling identification result to obtain multiple sub-areas, including: Voronoi segmentation is performed based on the virtual boundary and the seedling points within the virtual boundary to obtain the multiple sub-areas, and the seedling points are the coordinate points of a single target plant seedling in the target plot.

[0016] According to the plant seedling emergence quality assessment method based on remote sensing images provided by the present invention, the sub-regions are classified based on their areas to obtain the first sub-region, the second sub-region, and the third sub-region, including:

[0017] Normalizing the area of ​​each sub-region to obtain a normalized area of ​​each sub-region;

[0018] The sub-regions are classified based on the normalized areas to obtain the first sub-region, the second sub-region, and the third sub-region.

[0019] According to the plant seedling emergence quality assessment method based on remote sensing images provided by the present invention, determining the emergence quality assessment result of the target plant in the target plot based on the areas of the first sub-region, the second sub-region, and the third sub-region includes:

[0020] determining a reference number of seedlings based on the area of ​​the second sub-region and the area of ​​the target plot;

[0021] Determining the seedling deficiency rate based on the number of seedlings of all target plants and the number of reference seedlings in the plant seedling identification result;

[0022] The reseedling rate is determined based on the number of the third sub-area and the reference number of seedlings.

[0023] According to the plant seedling emergence quality assessment method based on remote sensing images provided by the present invention, determining the emergence quality assessment result of the target plant in the target plot based on the areas of the first sub-region, the second sub-region, and the third sub-region includes:

[0024] The uniformity of seedling emergence is determined based on the area of ​​each sub-region, the number of seedlings of all target plants in the plant seedling identification result, the average area of ​​all sub-regions, and the area variance of all sub-regions.

[0025] The present invention also provides a plant seedling quality assessment device based on remote sensing images, comprising:

[0026] A seedling identification module is configured to obtain a remote sensing image of a target plot, input the remote sensing image into a trained target detection model, and obtain a plant seedling identification result output by the target detection model, wherein the plant seedling identification result includes the coordinates of the target plant seedlings in the target plot in the remote sensing image;

[0027] an area determination module, configured to determine a seedling emergence area in the remote sensing image based on the plant seedling identification result, wherein the seedling emergence area is an area in the remote sensing image that includes the seedlings of the target plant in the target plot;

[0028] a region segmentation module, configured to perform Voronoi segmentation on the seedling emergence region in the remote sensing image based on the plant seedling identification result to obtain a plurality of subregions;

[0029] a region classification module, configured to classify the sub-regions based on their areas to obtain a first sub-region, a second sub-region, and a third sub-region, wherein the area of ​​the first sub-region is smaller than that of the second sub-region, and the area of ​​the second sub-region is smaller than that of the third sub-region;

[0030] A seedling quality assessment module is used to determine the seedling quality assessment result of the target plant in the target plot based on the areas of the first sub-area, the second sub-area and the third sub-area. The seedling quality assessment result reflects the seedling quality of the target plant in the target plot. The seedling quality assessment result includes the seedling deficiency rate and re-seedling rate, seedling uniformity and seedling vitality of the target crop in the target plot.

[0031] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the plant seedling emergence quality assessment method based on remote sensing images as described above is implemented.

[0032] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for assessing plant seedling emergence quality based on remote sensing images as described above is implemented.

[0033] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned methods for assessing plant seedling emergence quality based on remote sensing images.

[0034] The present invention provides a plant seedling emergence quality assessment method and device based on remote sensing images. The method and device process the remote sensing image of the target plot, first obtain the plant seedling identification result including the coordinates of the seedlings of the target plant in the target plot through the target detection model, then determine the area including the seedlings of the target plant in the target plot in the remote sensing image based on the plant seedling identification result, perform Voronoi segmentation on the area to obtain multiple sub-areas, and obtain the seedling quality assessment result reflecting the emergence quality of the target plant in the target plot based on the area of ​​the sub-area. No human participation in the seedling quality assessment is required throughout the process, and automated assessment of the seedling quality can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1 This is a flow chart of a method for assessing plant seedling quality based on remote sensing images provided by the present invention;

[0037] Figure 2 This is a schematic diagram of the plant seedling identification results in a plant seedling emergence quality assessment method based on remote sensing images provided by the present invention. Figure 1 ;

[0038] Figure 3 This is a schematic diagram of the plant seedling identification results in a plant seedling emergence quality assessment method based on remote sensing images provided by the present invention. Figure 2 ;

[0039] Figure 4 This is a schematic diagram of the performance results of corn seedling identification using the YoloV8 model and other models in a plant seedling emergence quality assessment method based on remote sensing images provided by the present invention. Figure 1 ;

[0040] Figure 5 This is a schematic diagram of the performance results of corn seedling identification using the YoloV8 model and other models in a plant seedling emergence quality assessment method based on remote sensing images provided by the present invention. Figure 2 ;

[0041] Figure 6 This is a schematic diagram of the Voronoi segmentation process in a plant seedling quality assessment method based on remote sensing images provided by the present invention. Figure 1 ;

[0042] Figure 7 This is a schematic diagram of the Voronoi segmentation process in a plant seedling quality assessment method based on remote sensing images provided by the present invention. Figure 2 ;

[0043] Figure 8 This is a schematic diagram of the Voronoi segmentation process in a plant seedling quality assessment method based on remote sensing images provided by the present invention. Figure 3 ;

[0044] Figure 9 This is a schematic diagram of the Voronoi segmentation process in a plant seedling quality assessment method based on remote sensing images provided by the present invention. Figure 4 ;

[0045] Figure 10 This is a schematic diagram of the Voronoi segmentation process in a plant seedling quality assessment method based on remote sensing images provided by the present invention. Figure 5 ;

[0046] Figure 11 This is a schematic diagram of the Voronoi segmentation process in a plant seedling quality assessment method based on remote sensing images provided by the present invention. Figure 6 ;

[0047] Figure 12 This is a diagram showing the result of identifying the missing seedling rate in an application example of a plant seedling emergence quality assessment method based on remote sensing images provided by the present invention;

[0048] Figure 13 This is a diagram showing the reseedling rate identification result in an application example of a plant seedling emergence quality assessment method based on remote sensing images provided by the present invention;

[0049] Figure 14This is a seedling emergence uniformity recognition result diagram in an application example of a plant seedling emergence quality assessment method based on remote sensing images provided by the present invention;

[0050] Figure 15 This is a schematic structural diagram of a plant seedling emergence quality assessment device based on remote sensing images provided by the present invention;

[0051] Figure 16 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0052] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0053] like Figure 1 As shown, the present invention provides a method for evaluating plant seedling quality based on remote sensing images, comprising the steps of:

[0054] S110, obtaining a remote sensing image of the target plot, inputting the remote sensing image into a trained target detection model, and obtaining a plant seedling recognition result output by the target detection model, wherein the plant seedling recognition result includes the coordinates of the target plant seedlings in the target plot in the remote sensing image;

[0055] S120, determining a seedling emergence area in the remote sensing image based on the plant seedling recognition result, where the seedling emergence area is a plot including seedlings of target plants in the target plot;

[0056] S130, performing Voronoi segmentation on the seedling emergence area in the remote sensing image based on the plant seedling identification result to obtain a plurality of sub-regions; S140, classifying the sub-regions based on their areas to obtain a first sub-region, a second sub-region, and a third sub-region, wherein the area of ​​the first sub-region is smaller than that of the second sub-region, and the area of ​​the second sub-region is smaller than that of the third sub-region;

[0057] S150. Determine the seedling quality assessment result of the target plant in the target plot based on the areas of the first sub-region, the second sub-region, and the third sub-region. The seedling quality assessment result reflects the seedling quality of the target plant in the target plot. The seedling quality assessment result includes the seedling deficiency rate and reseedling rate, seedling uniformity, and seedling vitality of the target crop in the target plot.

[0058] The method provided by the present invention processes the remote sensing image of the target plot, first obtains the plant seedling recognition result including the coordinates of the seedlings of the target plant in the target plot through the target detection model, then determines the area including the seedlings of the target plant in the target plot in the remote sensing image based on the plant seedling recognition result, performs Voronoi segmentation on the area to obtain multiple sub-areas, and obtains the seedling quality assessment result reflecting the seedling quality of the target plant in the target plot based on the area of ​​the sub-area. No manual participation in the seedling quality assessment is required throughout the process, and automatic assessment of the seedling quality can be achieved.

[0059] The method provided by the present invention can be used to evaluate the emergence quality of different crops such as corn, rice, soybean, cotton, etc. In other words, the target crops can be different crops such as corn, rice, soybean, cotton, etc.

[0060] The target plot is the plot where the seedling emergence quality of the target plant needs to be assessed. Remote sensing images of the target plot can be generated through various channels, such as remote sensing satellites or drone-mounted image generation equipment. To improve target detection accuracy, it is preferred to fly a drone equipped with a spectral sensor over the target plot to generate remote sensing images of the target plot. Specifically, fixed-wing drones can be used for remote sensing image generation for large areas of field crops, while conventional rotary-wing drones can be used for remote sensing image generation for smaller breeding areas. The drone-mounted sensors can be multispectral sensors with a wide viewing angle and wide bandwidth, which offer both high spatial resolution and spectral advantages. Depending on the target plant for seedling emergence quality assessment, the drone's flight altitude can be set to a different level. Generally, the optimal flight altitude is one where the ground resolution (GSD) of the captured images is better than 5 cm. The flight altitude is determined based on the seedling size of the plant. The flight altitude should be high enough for the human eye to distinguish the target plant seedlings from the remote sensing images. Drones are best flown on cloudy, windless days to avoid shadows cast by direct sunlight and the effects of wind on seedlings. Remote sensing images should be collected during the seedling stage of the target crop, when the seedlings have emerged but are not intersecting each other. For example, for corn, the optimal time for remote sensing image acquisition is between the V2-V5 stage. After the initial drone image acquisition, pre-processing such as image stitching can be performed using drone data processing tools such as Pix4D to generate remote sensing images for subsequent identification of target crop seedlings.

[0061] After obtaining the remote sensing image, the remote sensing image is input into the trained target detection model to obtain the plant seedling recognition results output by the target detection model. The target detection model can use an existing model for detecting specified targets. However, due to the complex environment within the farmland plot, the seedling recognition accuracy is insufficient in complex scenes. In one possible implementation, the Yolo V8 deep learning model is used as the target detection model for seedling recognition and coordinate extraction. The Yolo V8 deep learning model has better performance than models such as Faster-RCNN and Yolov5-s, as shown in Table 1. Figure 4 and Figure 5 As shown in Figure 2, the highest scores were achieved in terms of precision, recall, F1 score, and mAP@0.5. The actual verification was conducted using corn as the target plant, and the results of plant seedling recognition are shown in Figure 2. Figure 2 and Figure 3 As shown, Figure 2 This is the test result of corn seedlings in stage V3 on a sunny day. Figure 3 This is the detection result of corn seedlings in stage V2 on a cloudy day. It can be seen that the Yolo V8 deep learning model is also well adaptable to scenes with high overlap between seedlings and strong shadows.

[0062] Table 1

[0063]

[0064] Based on the plant seedling identification results obtained above, the seedling emergence area can be determined in the remote sensing image. Specifically, determining the seedling emergence area in the remote sensing image based on the plant seedling identification results includes:

[0065] Determine an initial plot in the remote sensing image based on the plant seedling recognition result, where the initial plot is a minimum convex polygonal plot that includes all seedlings of target plants in the target plot;

[0066] Expand the vertex coordinates of the initial plot to obtain the vertices of the seedling emergence area;

[0067] Insert virtual points between adjacent vertices of the emergence area to obtain the virtual boundary of the emergence area.

[0068] like Figure 6 and Figure 7 As shown in FIG, the initial plot is the smallest convex polygonal plot including the seedlings of all target plants in the target plot. In a possible implementation, the initial plot can be directly segmented as the seedling emergence area in the remote sensing image. However, the boundary effect will affect the Voronoi segmentation accuracy. The method provided by the present invention, after obtaining the initial plot, is as follows: Figure 8 As shown, the initial plot was expanded to obtain the seedling emergence area.

[0069] The initial plot may be expanded by expanding the convex polygon vertices corresponding to the initial plot to obtain vertices of the emergence area, and then inserting virtual points between adjacent vertices of the emergence area to obtain a virtual boundary of the emergence area.

[0070] The expansion process of the initial vertex is to move the vertex coordinates of the initial vertex in the direction of increasing the initial plot, such as Figure 8 In one possible implementation, the x-coordinate is expanded N times, and the y-coordinate is expanded M times, where N=1+2 / (n-1) and M=1+2 / (m-1), and n and m are the number of rows and columns of target crop plants in the target plot, respectively.

[0071] Virtual points are inserted at adjacent vertices in the seedling emergence region. These points can be inserted based on the average row and column spacing of the target crop within the target plot. By expanding the initial plot and adding virtual boundary points, the influence of boundary effects on the Voronoi segmentation can be eliminated.

[0072] Based on the plant seedling recognition results, Voronoi segmentation is performed on the seedling emergence area in the remote sensing image to obtain multiple sub-areas, including:

[0073] Voronoi segmentation is performed based on the virtual boundary and the seedling points within the virtual boundary to obtain multiple sub-regions. The seedling point is the coordinate point of a single target plant seedling in the target plot.

[0074] Voronoi segmentation can divide a region into multiple sub-regions, each of which contains only one target point. In the method provided by the present invention, the coordinate point of a single target plant seedling obtained based on target detection coordinate recognition is used as the target point for Voronoi segmentation to obtain multiple sub-regions. In other words, each sub-region contains only one target plant seedling. The process of dividing a region into multiple sub-regions is as follows: Figure 9 、 Figure 10 and Figure 11 As shown in the figure, first construct the Delaunay triangulation, then calculate the center of the circumscribed circle of each triangle, and connect all the centers to obtain each sub-area.

[0075] The area s of each sub-region i Based on this, the emergence quality assessment results of the target plants in the target plots are determined.

[0076] In the existing technology, the seedling quality assessment index is relatively simple, and the selection of indicators is not comprehensive, which is not enough to fully and precisely reflect the seedling quality. The assessment of each single indicator is not accurate enough, and the accuracy needs to be improved. For example, the seedling quantity index, the current method is to calculate the seedling rate or the seedling absence rate based on the number of seedlings identified divided by the total number of seedlings calculated based on the expected planting density of the plot. This method is limited in two aspects. On the one hand, it is limited by the accuracy of existing seedling detection and is difficult to adapt to complex environmental conditions (such as the situation where two or three seedlings emerge from a single hole, weeds, image shadows, etc.). There are many cases of missed detection or multiple detection. On the other hand, some scenes lack planting density data. For example, scenes with manual sowing rather than machine sowing are difficult to evaluate using this method. In addition, re-seedling refers to the situation where more than one seedling emerges from a single hole. The existing technology has also always lacked a better means of evaluating the re-seedling rate.

[0077] The first limitation of the emergence rate or the seedling missing rate can be improved by adopting the Yolo V8 target detection model. The second limitation of the emergence rate or the seedling missing rate and the lack of better means for evaluating the re-seedling rate can be improved by the method provided by the present invention by classifying the sub-regions according to their areas.

[0078] In the method provided by the present invention, the seedling quality assessment result includes the seedling shortage rate and seedling recovery rate of the target crop in the target plot. As described above, each sub-region obtained after Voronoi segmentation of the seedling emergence area only includes the seedlings of one target plant. Therefore, according to the size of the sub-region, it can be confirmed what type of area the sub-region corresponds to. For larger sub-regions, it is very likely that there is a seedling shortage phenomenon, and for smaller sub-regions, it is very likely that there is a seedling recovery phenomenon. In the method provided by the present invention, the sub-regions are divided into three categories based on the area of ​​the sub-regions: a first sub-region (corresponding to the seedling recovery area), a second sub-region (for the normal seedling emergence area) and a third sub-region (corresponding to the seedling shortage area).

[0079] Furthermore, when classifying sub-regions based on their areas, in order to avoid the influence of magnitude, in a possible implementation, the areas of all sub-regions may be normalized, and classification may be performed based on the normalized areas.

[0080] The process of normalizing the area of ​​the sub-region can be expressed as:

[0081]

[0082] s i is the area of ​​the i-th sub-region; S min is the minimum area in the subregion; S max is the maximum area among all sub-regions.

[0083] Normalized area s * The smaller sub-region is the first sub-region, which is considered as the re-seedling region; the larger sub-region is the third sub-region, which is considered as the seedling-deficient region; the moderate sub-region is the second sub-region, which is the normal seedling region. Specifically, the sub-regions are classified according to the relationship between the normalized area size of the sub-region and the preset first threshold θ1 and second threshold θ2. If s * ≥θ1, it is a missing seedling unit; if s * ≤θ2, it is a replanting unit; if θ2 * <θ1, it is a normal unit. The thresholds θ1 and θ2 are set according to the distribution of seedlings in the target plot. They can be manually determined or clustered according to the area distribution of each sub-region in the target plot.

[0084] The seedling shortage rate and seedling recovery rate are determined based on the areas of the first sub-area, the second sub-area, and the third sub-area, including:

[0085] determining a reference number of seedlings based on the area of ​​the second sub-region and the area of ​​the target plot;

[0086] Determine the seedling deficiency rate based on the number of seedlings of all target plants and the number of reference seedlings in the plant seedling identification results;

[0087] The reseedling rate was determined based on the number of the third sub-area and the reference number of seedlings.

[0088] The seedling loss rate and reseedling rate of target plants in the target plot can be calculated based on the following formula:

[0089]

[0090] Among them, G is the seedling loss rate, R is the seedling recovery rate, r is the number determined to be the first sub-area, n is the total number of seedlings in the target plot, m is the average area of ​​each second sub-area, S is the total area of ​​the target plot, and n′ is the total area of ​​the target plot divided by the average area of ​​the second sub-area, that is, the total number of seedlings in the target plot under ideal distribution.

[0091] like Figure 12 and Figure 13 As shown, Figure 12 and Figure 13 These are the results of identifying the seedling-missing rate and the seedling-recovery rate of different corn varieties using the method provided by the present invention.

[0092] ​In one possible implementation, in response to the defect of a single crop seedling quality assessment index in the prior art, in the method provided by the present invention, the seedling quality assessment result also includes the seedling vigor of the target plant in the target plot. In the prior art, the seedling vigor index is judged solely by NDVI (normalized difference vegetation index), which is not comprehensive. For example, due to variety differences, NDVI cannot fully reflect the seedling vigor of different varieties. In response to this defect, in one implementation of the present invention, the seedling vigor index includes the seedling vigor of a single seedling and the seedling vigor of the group as a whole in the target plot. The specific calculation formula is:

[0093]

[0094] Among them, G rowthi is the seedling strength of a single seedling, NDVI i is the NDVI of a single seedling, A reai is the green leaf coverage area of ​​a single seedling in the target plot; is the overall seedling potential of the current regional population, and n is the total number of seedlings in the target plot.

[0095] Furthermore, the seedling quality assessment results can also include seedling uniformity. The traditional plant spacing assessment method is difficult to accurately reflect the actual spatial distribution of seedlings, because the one-dimensional indicator of plant spacing is difficult to accurately describe the two-dimensional spatial distribution uniformity of crops. In addition, the current uniformity assessment is only an assessment of the position distribution uniformity, and it is difficult to accurately assess the growth distribution uniformity of individual plants and the area distribution uniformity of individual plants.

[0096] To address this shortcoming, in one implementation of the method provided by the present invention, determining the seedling quality assessment result of the target plant in the target plot based on the area of ​​the sub-region includes:

[0097] The uniformity of seedling emergence was determined based on the area of ​​each sub-region, the number of seedlings of all target plants in the plant seedling identification results, the average area of ​​all sub-regions, and the area variance of all sub-regions.

[0098] Seedling emergence uniformity can include position distribution uniformity VPUI, single plant growth distribution uniformity and single plant area distribution uniformity. The higher the position distribution uniformity VPUI, the more uniform the spatial distribution of seedlings. 0 represents extremely uneven, and 1 represents completely uniform. Its calculation formula is as follows:

[0099]

[0100] Where: VPUI is the position distribution uniformity (Voronoi Polygon Uniformity Index), s iis the area of ​​the sub-region corresponding to the i-th seedling, n is the total number of seedlings in the target plot, is the average of all sub-region areas, and σ is the variance of all sub-region areas.

[0101] like Figure 14 As shown, Figure 14 This is a diagram showing the results of identifying the uniformity of position distribution of different corn varieties using the method provided by the present invention.

[0102] The calculation method for the uniformity of the growth distribution of a single plant and the uniformity of the area distribution of a single plant is similar to the calculation formula for the position distribution, but is based on different Voronoi segmentation results. That is, in one possible implementation, Voronoi segmentation is performed on the seedling emergence area in the remote sensing image based on the plant seedling identification result to obtain multiple sub-areas, including: performing Voronoi segmentation on the seedling emergence area based on the first weight sequence and the plant seedling identification result to obtain multiple first weight sub-areas;

[0103] Performing Voronoi segmentation on the seedling emergence area based on the second weight sequence and the plant seedling identification result to obtain a plurality of second weight sub-areas;

[0104] Voronoi segmentation is performed on the seedling emergence area based on the third weight sequence and the plant seedling identification result to obtain a plurality of third weight sub-areas.

[0105] The first weight sequence, the second weight sequence, and the third weight sequence all include weight values ​​for the coordinate points corresponding to the seedlings of each target plant in the target plot. The weight values ​​for the coordinate points corresponding to the seedlings of each target plant included in the first weight sequence can all be 1, the weight values ​​for the coordinate points corresponding to the seedlings of each target plant included in the second weight sequence are determined based on the NDVI of each seedling, and the weight values ​​for the coordinate points corresponding to the seedlings of each target plant included in the third weight sequence are determined based on the green leaf coverage area of ​​each seedling.

[0106] When calculating the emergence rate, re-seedling rate and position distribution uniformity, the results obtained by classifying each first-weighted sub-region are used. When calculating the uniformity of single plant growth distribution, the results obtained by classifying each second-weighted sub-region are used, and the same formula as the position distribution uniformity is used for calculation. When calculating the uniformity of single plant area distribution, the results obtained by classifying each third-weighted sub-region are used, and the same formula as the position distribution uniformity is used for calculation.

[0107] Based on the results of seedling emergence quality assessment, the quality of crop seedling emergence can be quantitatively described from indicators of multiple dimensions. By setting multiple indicators, the problem that the seedling emergence quality evaluation indicators in the existing technology are single, one-sided and have low accuracy can be solved.

[0108] The plant seedling quality assessment device based on remote sensing images provided by the present invention is described below. The plant seedling quality assessment device based on remote sensing images described below and the plant seedling quality assessment method based on remote sensing images described above can be used for reference. Figure 15 As shown, the plant seedling quality assessment device based on remote sensing images provided by the present invention includes:

[0109] The seedling identification module 1510 is configured to obtain a remote sensing image of a target plot, input the remote sensing image into a trained target detection model, and obtain a plant seedling identification result output by the target detection model. The plant seedling identification result includes the coordinates of the target plant seedlings in the target plot in the remote sensing image.

[0110] The region determination module 1520 is configured to determine a seedling emergence region in the remote sensing image based on the plant seedling identification result, where the seedling emergence region is a region in the remote sensing image that includes seedlings of target plants in the target plot;

[0111] A region segmentation module 1530 is configured to perform Voronoi segmentation on the seedling emergence region in the remote sensing image based on the plant seedling recognition result to obtain a plurality of sub-regions;

[0112] a region classification module 1540 configured to classify the subregions based on their areas to obtain a first subregion, a second subregion, and a third subregion, wherein the area of ​​the first subregion is smaller than that of the second subregion, and the area of ​​the second subregion is smaller than that of the third subregion;

[0113] The seedling quality assessment module 1550 is used to determine the seedling quality assessment result of the target plant in the target plot based on the areas of the first sub-area, the second sub-area and the third sub-area. The seedling quality assessment result reflects the seedling quality of the target plant in the target plot. The seedling quality assessment result includes the seedling deficiency rate and re-seedling rate, seedling uniformity and seedling vitality of the target crop in the target plot.

[0114] Figure 16 An example of a physical structure diagram of an electronic device is shown below. Figure 16As shown, the electronic device may include: a processor (processor) 1610, a communication interface (Communications Interface) 1620, a memory (memory) 1630 and a communication bus 1640, wherein the processor 1610, the communication interface 1620, and the memory 1630 communicate with each other through the communication bus 1640. The processor 1610 can call the logic instructions in the memory 1630 to execute the plant seedling emergence quality assessment method based on remote sensing images. The plant seedling emergence quality assessment method based on remote sensing images includes: obtaining a remote sensing image of a target plot, inputting the remote sensing image into a trained target detection model, obtaining a plant seedling recognition result output by the target detection model, the plant seedling recognition result including the coordinates of the seedlings of the target plant in the target plot in the remote sensing image; determining the seedling emergence area in the remote sensing image based on the plant seedling recognition result, the seedling emergence area being the plot including the seedlings of the target plant in the target plot; and Voronoi segmentation is performed on the seedling emergence area in the remote sensing image to obtain multiple sub-areas; the sub-areas are classified based on their areas to obtain a first sub-area, a second sub-area and a third sub-area, the area of ​​the first sub-area is smaller than that of the second sub-area, and the area of ​​the second sub-area is smaller than that of the third sub-area; based on the areas of the first sub-area, the second sub-area and the third sub-area, a seedling emergence quality assessment result of a target plant in a target plot is determined, the seedling emergence quality assessment result reflects the seedling emergence quality of the target plant in the target plot, and the seedling emergence quality assessment result includes the seedling deficiency rate and the seedling restoration rate, the seedling emergence uniformity and the seedling vitality of the target crop in the target plot.

[0115] In addition, the logic instructions in the above-mentioned memory 1630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0116] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the above-mentioned plant seedling emergence quality assessment method based on remote sensing images. The plant seedling emergence quality assessment method based on remote sensing images includes: obtaining a remote sensing image of a target plot, inputting the remote sensing image into a trained target detection model, obtaining a plant seedling recognition result output by the target detection model, the plant seedling recognition result including the coordinates of the seedlings of the target plant in the target plot in the remote sensing image; determining the seedling emergence area in the remote sensing image based on the plant seedling recognition result, and the seedling emergence area The invention relates to a plot including seedlings of target plants in a target plot; performing Voronoi segmentation on the seedling emergence area in the remote sensing image based on the plant seedling recognition result to obtain multiple sub-regions; classifying the sub-regions based on the areas of the sub-regions to obtain a first sub-region, a second sub-region and a third sub-region, wherein the area of ​​the first sub-region is smaller than that of the second sub-region, and the area of ​​the second sub-region is smaller than that of the third sub-region; determining the seedling emergence quality assessment result of the target plant in the target plot based on the areas of the first sub-region, the second sub-region and the third sub-region, the seedling emergence quality assessment result reflects the seedling emergence quality of the target plant in the target plot, and the seedling emergence quality assessment result includes the seedling deficiency rate and the seedling restoration rate, the seedling emergence uniformity and the seedling vitality of the target crop in the target plot.

[0117] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute the above-mentioned plant seedling emergence quality assessment method based on remote sensing images, and the plant seedling emergence quality assessment method based on remote sensing images comprises: obtaining a remote sensing image of a target plot, inputting the remote sensing image into a trained target detection model, obtaining a plant seedling recognition result output by the target detection model, the plant seedling recognition result including the coordinates of the seedlings of the target plant in the target plot in the remote sensing image; determining a seedling emergence area in the remote sensing image based on the plant seedling recognition result, the seedling emergence area being the seedlings including the target plant in the target plot. The invention relates to a method for determining a seedling emergence quality assessment result of a target plant in a target plot; performing Voronoi segmentation on the seedling emergence area in the remote sensing image based on the plant seedling recognition result to obtain multiple sub-areas; classifying the sub-areas based on the area of ​​the sub-areas to obtain a first sub-area, a second sub-area and a third sub-area, wherein the area of ​​the first sub-area is smaller than that of the second sub-area, and the area of ​​the second sub-area is smaller than that of the third sub-area; determining a seedling emergence quality assessment result of a target plant in a target plot based on the areas of the first sub-area, the second sub-area and the third sub-area, wherein the seedling emergence quality assessment result reflects the seedling emergence quality of the target plant in the target plot, and the seedling emergence quality assessment result includes the seedling deficiency rate and the seedling restoration rate, the seedling emergence uniformity and the seedling vitality of the target crop in the target plot.

[0118] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0119] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0120] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for evaluating plant seedling quality based on remote sensing images, characterized in that: include: Acquire a remote sensing image of a target plot, input the remote sensing image into a trained target detection model, and acquire a plant seedling recognition result output by the target detection model, wherein the plant seedling recognition result includes the coordinates of the seedlings of the target plants in the target plot in the remote sensing image; Determine a seedling emergence area in the remote sensing image based on the plant seedling identification result, wherein the seedling emergence area is an area in the remote sensing image that includes the seedlings of the target plant in the target plot; Performing Voronoi segmentation on the seedling emergence area in the remote sensing image based on the plant seedling recognition result to obtain a plurality of sub-areas; Classifying the sub-regions based on their areas to obtain a first sub-region, a second sub-region, and a third sub-region, wherein the area of ​​the first sub-region is smaller than that of the second sub-region, and the area of ​​the second sub-region is smaller than that of the third sub-region; Determine a seedling quality assessment result of the target plant in the target plot based on the areas of the first sub-region, the second sub-region, and the third sub-region, wherein the seedling quality assessment result reflects the seedling quality of the target plant in the target plot, and the seedling quality assessment result includes a seedling missing rate and a seedling recovery rate, seedling uniformity, and seedling vigor of the target plant in the target plot; The classifying the sub-regions based on the areas of the sub-regions to obtain a first sub-region, a second sub-region and a third sub-region comprises: Normalizing the area of ​​each sub-region to obtain a normalized area of ​​each sub-region; Classifying the sub-regions based on the normalized areas to obtain the first sub-region, the second sub-region, and the third sub-region specifically includes: The sub-regions are classified according to the size relationship between the normalized area of ​​the sub-region and the preset first threshold θ1 and second threshold θ2. * ≥θ1, is the third sub-region; if s * ≤θ2, is the first sub-region; if θ2 * <θ1, is the second sub-region.​ 2. The method for evaluating plant seedling quality based on remote sensing images according to claim 1, characterized in that: The step of determining the seedling emergence area in the remote sensing image based on the plant seedling identification result includes: Determine an initial plot in the remote sensing image based on the plant seedling recognition result, wherein the initial plot is a minimum convex polygonal plot including all the seedlings of the target plant in the target plot; Expanding the vertex coordinates of the initial plot to obtain the vertices of the seedling emergence area; Virtual points are inserted between adjacent vertices of the seedling emergence area to obtain a virtual boundary of the seedling emergence area.

3. The method for evaluating plant seedling quality based on remote sensing images according to claim 2, characterized in that: The step of performing Voronoi segmentation on the seedling emergence area in the remote sensing image based on the plant seedling identification result to obtain a plurality of sub-areas includes: Voronoi segmentation is performed based on the virtual boundary and the seedling points within the virtual boundary to obtain the multiple sub-areas, and the seedling points are coordinate points of a single target plant seedling in the target plot.

4. The method for evaluating plant seedling quality based on remote sensing images according to claim 1, characterized in that: The step of determining the seedling quality assessment result of the target plant in the target plot based on the areas of the first sub-region, the second sub-region, and the third sub-region includes: determining a reference number of seedlings based on the area of ​​the second sub-region and the area of ​​the target plot; Determining the seedling deficiency rate based on the number of seedlings of all the target plants in the plant seedling identification result and the number of reference seedlings; The re-seedling rate is determined based on the number of the third sub-area and the reference number of seedlings.

5. The method for evaluating plant seedling quality based on remote sensing images according to claim 1, characterized in that: The step of determining the seedling quality assessment result of the target plant in the target plot based on the areas of the first sub-region, the second sub-region, and the third sub-region includes: The uniformity of seedling emergence is determined based on the area of ​​each of the sub-regions, the number of seedlings of all the target plants in the plant seedling identification result, the average area of ​​all the sub-regions, and the variance of the area of ​​all the sub-regions.

6. A plant seedling quality assessment device based on remote sensing images, characterized in that: include: A seedling identification module is used to obtain a remote sensing image of a target plot, input the remote sensing image into a trained target detection model, and obtain a plant seedling identification result output by the target detection model, wherein the plant seedling identification result includes the coordinates of the seedlings of the target plant in the target plot in the remote sensing image; A region determination module, configured to determine a seedling emergence region in the remote sensing image based on the plant seedling identification result, wherein the seedling emergence region is a region in the remote sensing image that includes the seedlings of the target plant in the target plot; A region segmentation module, configured to perform Voronoi segmentation on the seedling emergence region in the remote sensing image based on the plant seedling identification result to obtain a plurality of sub-regions; a region classification module, configured to classify the sub-regions based on their areas to obtain a first sub-region, a second sub-region, and a third sub-region, wherein the area of ​​the first sub-region is smaller than that of the second sub-region, and the area of ​​the second sub-region is smaller than that of the third sub-region; A seedling quality assessment module, used to determine the seedling quality assessment result of the target plant in the target plot based on the areas of the first sub-region, the second sub-region and the third sub-region, wherein the seedling quality assessment result reflects the seedling quality of the target plant in the target plot, and the seedling quality assessment result includes the seedling missing rate and the seedling recovery rate, the seedling uniformity and the seedling vigor of the target plant in the target plot; The classifying the sub-regions based on the areas of the sub-regions to obtain a first sub-region, a second sub-region and a third sub-region comprises: Normalizing the area of ​​each sub-region to obtain a normalized area of ​​each sub-region; Classifying the sub-regions based on the normalized areas to obtain the first sub-region, the second sub-region, and the third sub-region specifically includes: The sub-regions are classified according to the size relationship between the normalized area of ​​the sub-region and the preset first threshold θ1 and second threshold θ2. * ≥θ1, is the third sub-region; if s * ≤θ2, is the first sub-region; if θ2 * <θ1, is the second sub-region.​ 7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the plant seedling emergence quality assessment method based on remote sensing images as described in any one of claims 1 to 5 is implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for assessing plant seedling quality based on remote sensing images according to any one of claims 1 to 5 is implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for assessing plant seedling quality based on remote sensing images according to any one of claims 1 to 5 is implemented.

Images