CN117011316B - Method and system for identifying internal structure of soybean stalk based on CT image - Google Patents
Method and system for identifying internal structure of soybean stalk based on CT image Download PDFInfo
- Publication number
- CN117011316B CN117011316B CN202311278665.4A CN202311278665A CN117011316B CN 117011316 B CN117011316 B CN 117011316B CN 202311278665 A CN202311278665 A CN 202311278665A CN 117011316 B CN117011316 B CN 117011316B
- Authority
- CN
- China
- Prior art keywords
- image
- resolution
- data
- soybean
- noise
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 244000068988 Glycine max Species 0.000 title claims abstract description 70
- 235000010469 Glycine max Nutrition 0.000 title claims abstract description 62
- 238000000034 method Methods 0.000 title claims abstract description 59
- 241000196324 Embryophyta Species 0.000 claims abstract description 44
- 238000012545 processing Methods 0.000 claims abstract description 29
- 238000009792 diffusion process Methods 0.000 claims abstract description 27
- 230000011218 segmentation Effects 0.000 claims abstract description 18
- 238000010603 microCT Methods 0.000 claims abstract description 6
- 238000012549 training Methods 0.000 claims description 29
- 238000005070 sampling Methods 0.000 claims description 17
- 238000003860 storage Methods 0.000 claims description 15
- 238000003709 image segmentation Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 11
- 238000013434 data augmentation Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000012952 Resampling Methods 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 claims description 3
- 230000036544 posture Effects 0.000 claims description 3
- 238000012887 quadratic function Methods 0.000 claims description 3
- 238000009827 uniform distribution Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 6
- 230000001066 destructive effect Effects 0.000 abstract 1
- 238000002591 computed tomography Methods 0.000 description 83
- 238000004590 computer program Methods 0.000 description 9
- 238000009395 breeding Methods 0.000 description 4
- 230000001488 breeding effect Effects 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000002708 enhancing effect Effects 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 238000013144 data compression Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000009418 agronomic effect Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 229920001296 polysiloxane Polymers 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 239000010979 ruby Substances 0.000 description 1
- 229910001750 ruby Inorganic materials 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4053—Scaling of whole images or parts thereof, e.g. expanding or contracting based on super-resolution, i.e. the output image resolution being higher than the sensor resolution
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Geometry (AREA)
- Image Processing (AREA)
Abstract
A soybean stalk internal structure identification method and system based on CT image, the method includes: step one, placing soybean pot plants into plant CT acquisition equipment, and scanning soybean plants from top to bottom by utilizing micro-CT equipment to obtain acquired soybean stalk CT image data; step two, data enhancement is carried out on the acquired CT image data set, a diffusion model is introduced to carry out super-resolution processing on the data, and richer tissue detail information is obtained; dividing the CT image to obtain a division map of the tissue structure inside the soybean stalk; and step four, carrying out volume reconstruction according to the segmentation result to obtain gridding data and obtaining parameters such as the volume, the surface area and the like of the segmentation area. According to the invention, the internal tissues of the stems in the whole life cycle of soybeans are captured by utilizing a CT technology, so that the traditional destructive capture of internal information of soybean plants is avoided, the fine granularity of identifying the internal structures of the stems of the soybean plants can be effectively improved, and the accurate and fine phenotype identification is realized.
Description
Technical Field
The invention belongs to the field of computer vision and image processing, and relates to a soybean stalk internal structure identification method and system based on CT images.
Background
With the rapid development of advanced breeding equipment and computer technology, growing attention is currently being directed to phenotype driven computational breeding systems, genotype driven computational breeding systems, and artificial intelligence driven multi-module computational breeding systems. Genomic data is currently quite voluminous and corresponding phenotypic data is still relatively scarce. The phenotype data acquisition is rough, the accuracy and fineness are not enough, and less phenotype data cannot be well associated with huge genome data, so that corresponding refining equipment is required to acquire more fine phenotypes, and the two-dimensional three-dimensional phenotype group equipment such as hyperspectral, multispectral, near infrared, fluorescence, RGB, X-ray Computer Tomography (CT) and the like is gradually applied to the phenotype acquisition of soybeans, so that the information such as the full-growth period growth development and important agronomic characters of the soybeans can be acquired with high precision. Wherein CT is an excellent tool for 3D imaging of internal tissues and organs of plants, and can avoid damaging plant bodies, thereby realizing the observation of the whole growth cycle of plants. The advent of CT image data provides more refined and accurate internal phenotype data, which then also presents new challenges. CT images are slow in acquisition speed, high-magnification sampling can obtain high-resolution tissues, but the acquisition area is small, so that the CT images are acquired by common magnification, and a slice image with an ideal size can be obtained, but the resolution is lower. When the CT image data is rapidly acquired, the higher definition is maintained, and the key phenotype data is extracted from the CT image, so that a method suitable for processing the plant CT image needs to be researched, and the accurate, fine and efficient phenotype extraction is realized.
Disclosure of Invention
In order to solve the technical problems in the prior art, the invention provides a soybean stalk internal structure identification method and system based on CT images.
The method of the invention introduces the current latest denoising diffusion model into image super-resolution processing to acquire more abundant tissue detail information, then introduces an image segmentation algorithm based on medical CT images into plant CT image segmentation to acquire a segmentation map of tissue structures inside soybean stalks, and performs three-dimensional reconstruction according to segmentation results to acquire parameters such as volume, surface area and the like of a segmentation area, and the specific technical scheme is as follows:
the invention discloses a soybean stalk internal structure identification method based on CT images, which comprises the following steps:
(1) Placing the soybean pot in a plant CT acquisition device, and scanning soybean plants from top to bottom by utilizing micro-CT equipment to obtain acquired soybean stalk CT image data;
(2) The method comprises the steps of performing enhancement processing on a picture acquired in a common scanning mode, introducing a diffusion model into an image super-resolution network, inputting the acquired picture into a CT image super-resolution network based on the diffusion model for enhancement processing, and obtaining a high-resolution tissue image;
(3) Performing data augmentation treatment on the treated enhanced image, inputting the enhanced image into an image segmentation network for plant region segmentation to obtain a mask image, and performing pixel point multiplication on the mask image and a corresponding CT image to obtain an image of a plant region;
(4) And reconstructing the volume of the segmented CT image to obtain gridding data, and obtaining parameters such as the volume, the surface area and the like of the segmented region to accurately obtain the phenotype data inside the plant.
Further, the step (1) specifically includes the following substeps:
the method comprises the steps of (1.1) collecting training data, collecting plants by adopting common magnification and high-precision mode magnification respectively, and processing common magnification pictures and high-precision mode magnification pictures to obtain corresponding stalk structures, wherein the corresponding stalk structures are used as high-low resolution image pairs of the stalk structures;
(1.2) in the actual use stage, scanning plants to obtain a large-area acquisition image with lower resolution;
further, the step (2) specifically includes the following substeps:
and (2.1) performing super-resolution network training based on the denoising diffusion model. The low resolution CT input picture is LR, HR is the corresponding original high resolution CT picture, and the LR is subjected to a pre-training encoder model to obtain an initial prediction high resolution pictureInput into the Unet network. UP (LR) is a picture directly after bicubic upsampling for low resolution, data initial state during diffusion +.>The difference between the HR image and the UP-sampled picture UP (LR) contains high frequency information. The diffusion process can be directly based on +.>To arbitrary->Step->Sampling is carried out by the characteristic of the diffusion process, i.e. sampling time step +.>Variance of time->Actual sampling noise->Directly obtain->Of (1), wherein->Belonging to uniform distribution, i.e.，/>Wherein->WhereinThe variance used for each step is between 0 and 1. During training, will ∈ ->，/>And->Input the same to the Unet network to obtain estimated noise +.>And is +.>Comparing, calculating loss by estimating noiseTraining and optimizing the noise estimation network, minimizing a loss function, and finally obtaining estimated noise (I) meeting convergence conditions>。
And (2.2) performing super-resolution inference on the low-resolution image to obtain a high-resolution picture. Each step ofCan only be obtained from the previous step. Hidden variable obtained by passing LR image through Encoder +.>Adding a conditional noise predictor to guide the generation of corresponding HR information while initially randomly sampling the noise figure +.>And->As input to the Unet, the calculation gets the +.>Noise ∈estimated by step>Then add the randomly sampled perturbation ++>Calculated by the following formula->Is a noise figure of (1):
wherein the method comprises the steps ofIn this way, circulation is continued until +.>Finally->And adding the UP-sampled image UP (LR) with the LR to finally obtain a high-fraction CT image SR, wherein the high-fraction CT image SR is used for CT image enhancement under the condition of multiple acquisition of a common method.
Further, the step (3) specifically includes the following substeps:
(3.1) carrying out data augmentation processing on the processed high-resolution image, simulating different scanning parameters by using a contrast enhancement means, simulating different positioning postures (the maximum rotation angle is 45 degrees) during scanning by using random rotation, and adding random up-down overturning and normalization operations, wherein the normalization expression is as follows:
wherein the method comprises the steps ofFor inputting an image +.>For normalized image, ++>And->The maximum pixel value and the minimum pixel value of the input image, respectively.
(3.2) inputting the processed image into an open-source pre-training image segmentation network based on deep learning for training and optimizing so as to adapt to the CT image of the soybean stalk, and segmenting different tissues of the CT image of the soybean stalk through the optimized segmentation network;
further, the step (4) specifically includes the following substeps:
(4.1) collecting the segmented CT image area, obtaining three-dimensional data of soybean stalk tissues, resampling the interior of the three-dimensional data by using a quadratic function sampling mode, thereby realizing data compression, reducing the calculated amount in the surface reconstruction process, extracting an isosurface by using a Maring cube algorithm, calculating a surface normal, obtaining triangular patches, and splicing the triangular patches to form a continuous grid three-dimensional curved surface;
(4.2) calculating parameters of thickness, included angle, radius, surface area and volume of the meshed three-dimensional curved surface, obtaining phenotype parameters of the partitioned areas, and accurately obtaining phenotype data inside plants;
the other side of the invention comprises one or more processors for realizing the soybean stalk internal structure identification method based on the CT image.
Another aspect of the present invention includes a computer-readable storage medium having stored thereon a program which, when executed by a processor, implements the CT image-based soybean stalk internal structure identification method of the present invention.
Another side of the present invention is a soybean stalk internal structure recognition system based on CT images, comprising:
the soybean stalk CT image data acquisition module is used for placing soybean pot plants into plant CT acquisition equipment, and scanning soybean plants from top to bottom by utilizing micro-CT equipment to obtain acquired soybean stalk CT image data;
the image enhancement module is used for enhancing the acquired image in the common scanning mode, introducing the diffusion model into the image super-resolution network, inputting the acquired image into the CT image super-resolution network based on the diffusion model for enhancing, and obtaining a high-resolution tissue image;
the plant area image acquisition module is used for carrying out data augmentation processing on the processed enhanced image, inputting the enhanced image into the image segmentation network for plant area segmentation to obtain a mask image, and carrying out pixel point multiplication on the mask image and the corresponding CT image to obtain an image of a plant area;
and the plant internal phenotype data acquisition module is used for reconstructing the volume of the segmented CT image to obtain gridding data and accurately acquiring plant internal phenotype data.
The other side of the invention is a computing device which comprises a memory and a processor, wherein executable codes are stored in the memory, and the processor realizes the soybean stalk internal structure identification method based on the CT image when executing the executable codes.
Another aspect of the invention is a computer program product comprising a computer program which, when executed by a processor, implements a CT image based soybean stalk internal structure recognition method of the invention.
The beneficial effects of the invention are as follows: aiming at the characteristic that the information such as the grain edges of the soybean internal stalk images shot by CT is rich, a diffusion model is introduced to activate more pixels to reconstruct a high-resolution result, and the super-resolution technology recovers more detail information while the CT image acquisition area is larger, so that the further segmentation effect of various tissues is improved. By introducing a pre-training image segmentation network based on deep learning, the method is different from the traditional plant CT image segmentation modes based on a region growing method and the like, can extract characteristic information more effectively and realize tissue segmentation with higher precision. On the basis of the segmentation, the phenotype parameters are obtained through a reconstruction technology, so that the precision, refinement and intellectualization level of plant phenotype extraction are effectively improved.
Drawings
FIG. 1 is a schematic overall flow diagram of a soybean stalk internal structure identification method based on CT images;
FIG. 2 is a schematic flow chart of a super-resolution training process based on a diffusion model according to the invention;
FIG. 3 is a schematic flow chart of a super-resolution reasoning process based on a diffusion model of the invention;
FIG. 4 is a schematic view of the CT image segmentation, three-dimensional reconstruction and parameter calculation flow of the present invention;
FIG. 5 is a schematic structural diagram of a soybean stalk internal structure identification device based on CT images;
fig. 6 is a system configuration diagram of the present invention.
Detailed Description
In order to make the objects, technical solutions and technical effects of the present invention more apparent, the present invention will be further described in detail below with reference to the accompanying drawings.
Example 1
Examples: as shown in fig. 1, the soybean stalk internal structure identification method based on the CT image comprises the following steps:
(1) Placing the soybean pot in a plant CT acquisition device, and scanning soybean plants from top to bottom by utilizing micro-CT equipment to obtain acquired soybean stalk CT image data;
the method comprises the steps of (1.1) collecting training data, collecting plants by adopting common magnification and high-precision mode magnification respectively, and processing common magnification pictures and high-precision mode magnification pictures to obtain corresponding stalk structures, wherein the corresponding stalk structures are used as high-low resolution image pairs of the stalk structures;
(1.2) scanning plants to obtain a large-area acquisition image with lower resolution ratio during actual acquisition;
(2) The method comprises the steps of performing enhancement processing on a picture acquired in a common scanning mode, introducing a diffusion model into an image super-resolution network, inputting the acquired picture into a CT image super-resolution network based on the diffusion model for enhancement processing, and obtaining a high-resolution tissue image; the CT image super-resolution network takes a low-resolution image as input based on a denoising diffusion model, and then generates a high-resolution picture. The model is divided into two phases, one is a training phase, as shown in fig. 2, and the other is an reasoning phase, as shown in fig. 3, wherein a neural network is needed in the step of noise estimation, and a Unet model is adopted.
(2.1) the low resolution CT input picture is LR, HR is the corresponding original high resolution CT picture, and LR is obtained by a pre-training encoder modelInitial pre-prediction resolution pictureThe pre-training encoder model in this embodiment, which is input into the Unet network, uses a pre-trained residual intensive block (RRDB) structure, and the pre-training encoder uses an L1 loss function. UP (LR) is a picture directly after bicubic upsampling for low resolution, data initial state during diffusion +.>The difference between the HR image and the UP-sampled picture UP (LR) contains high frequency information. The diffusion process can be directly based on the residual image +.>To arbitrary->Step->Sampling is carried out by the characteristic of the diffusion process, i.e. sampling time step +.>Variance of timeActual sampling noise->Directly obtain->Of (1), wherein->Belonging to uniform distribution, i.e.)>，/>Wherein->Wherein->The variance used for each step is between 0 and 1. During training, will ∈ ->，/>And->Input the same to the Unet network to obtain estimated noise +.>And is +.>Comparing, calculating the loss by estimating the noise +.>Training and optimizing a noise estimation network, minimizing a loss function by using a gradient optimization algorithm, and finally obtaining estimated noise meeting convergence conditions>. Using PSNR (peak signal to noise ratio) and SSIM (structural similarity) as evaluation indexes of picture quality, selecting high-resolution medical CT image dataset for initial training data, and cutting soybean CT image dataset to ∈>And the pixels are taken as truth data, the truth data is subjected to degradation processing of bicubic downsampling, a CT image with high and low resolution is obtained, and in the embodiment, 8 times and 4 times downsampling is performed on resolution, so that the later 8 times and 4 times super resolution is adapted. High and low resolutionThe image pair is used as a training set, a verification set and a test set, the whole training process is divided into two stages, wherein the first stage adopts a processed medical CT image data set for training, the second stage adopts a processed soybean CT image data set for model adjustment, so that the image pair is more suitable for soybean tissues, a back propagation strategy is used for updating a network, and if the network is converged, a trained network model is saved for use as a final reasoning.
(2.2) super-resolution inference of LR images, each stepCan only be obtained from the previous step. Hidden variable obtained by passing LR image through Encoder +.>Adding a conditional noise predictor to guide the generation of corresponding HR information while initially randomly sampling the noise figure +.>And->As input to the Unet, the calculation gets the +.>Noise ∈estimated by step>Then add the randomly sampled perturbation ++>，/>When (I)>Obeys->Gaussian distribution->When=1,>. Calculated by the formula->Is a noise figure of (1):
wherein the method comprises the steps ofIn this way, circulation is continued until +.>Finally->And adding the UP-sampled image UP (LR) with the LR to finally obtain a high-fraction CT image SR, wherein the high-fraction CT image SR is used for CT image enhancement under the condition of multiple acquisition of a common method.
(3) Performing data augmentation treatment on the treated enhanced image, inputting the enhanced image into an image segmentation network for plant region segmentation to obtain a mask image, and performing pixel point multiplication on the mask image and a corresponding CT image to obtain an image of a plant region;
(3.1) carrying out data augmentation processing on the processed high-resolution image, wherein the contrast of the CT image is different under different scanning parameters, the contrast enhancement means is used for simulating different scanning parameters, and the random rotation is used for simulating different positioning postures (the maximum rotation angle is 45 degrees) during scanning, and in addition, random up-and-down overturning and normalization operations are added, so that the normalization expression is as follows:
wherein the method comprises the steps ofFor inputting an image +.>For normalized image, ++>And->The maximum pixel value and the minimum pixel value of the input image, respectively.
(3.2) inputting the processed soybean CT image into a pre-trained transformation-Unet image segmentation network for training optimization so as to adapt to the CT image of the soybean stalk, and performing different tissue segmentation of the soybean stalk CT image through the optimized segmentation network, as shown in figure 4;
(4) And carrying out volume reconstruction on the segmented CT image to obtain gridding data, and obtaining thickness, included angle, radius, surface area and volume parameters of the segmented region, wherein the phenotype data inside the plant is accurately obtained, as shown in figure 4.
(4.1) collecting the segmented CT image area, obtaining three-dimensional data of soybean stalk tissues, resampling the interior of the three-dimensional data by using a quadratic function sampling mode, thereby realizing data compression, reducing the calculated amount in the surface reconstruction process, extracting an isosurface by using a Maring cube algorithm, calculating a surface normal, obtaining triangular patches, and splicing the triangular patches to form a continuous grid three-dimensional curved surface;
(4.2) calculating parameters of thickness, included angle, radius, surface area and volume of the meshed three-dimensional curved surface, and obtaining phenotype parameters of the segmented areas, so as to accurately obtain phenotype data inside plants;
example 2
Corresponding to the embodiment of the soybean stalk internal structure identification method based on the CT image, the invention also provides an embodiment of the soybean stalk internal structure identification device based on the CT image.
Referring to fig. 5, the device for identifying the internal structure of the soybean stalk based on the CT image provided by the embodiment of the invention includes one or more processors for implementing the method for identifying the internal structure of the soybean stalk based on the CT image in the above embodiment.
The embodiment of the soybean stalk internal structure identification device based on the CT image can be applied to any equipment with data processing capability, and the equipment with data processing capability can be equipment or device such as a computer. The apparatus embodiments may be implemented by software, or may be implemented by hardware or a combination of hardware and software. Taking software implementation as an example, the device in a logic sense is formed by reading corresponding computer program instructions in a nonvolatile memory into a memory by a processor of any device with data processing capability. In terms of hardware, as shown in fig. 5, a hardware structure diagram of an apparatus with data processing capability according to the present invention, where the apparatus for identifying soybean stalk internal structure based on CT image is located, is shown in fig. 5, and in addition to a processor, a memory, a network interface, and a nonvolatile memory shown in fig. 5, any apparatus with data processing capability in the embodiment generally includes other hardware according to the actual function of the apparatus with data processing capability, which is not described herein.
The implementation process of the functions and roles of each unit in the above device is specifically shown in the implementation process of the corresponding steps in the above method, and will not be described herein again.
For the device embodiments, reference is made to the description of the method embodiments for the relevant points, since they essentially correspond to the method embodiments. The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purposes of the present invention. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
Example 3
The embodiment of the invention also provides a computer readable storage medium, on which a program is stored, which when executed by a processor, implements a soybean stalk internal structure identification method based on CT images in the above embodiment.
The computer readable storage medium may be an internal storage unit, such as a hard disk or a memory, of any of the data processing enabled devices described in any of the previous embodiments. The computer readable storage medium may also be an external storage device of the wind turbine generator, such as a plug-in hard disk, a Smart Media Card (SMC), an SD Card, a Flash memory Card (Flash Card), etc. provided on the device. Further, the computer readable storage medium may include both internal storage units and external storage devices of any data processing device. The computer readable storage medium is used for storing the computer program and other programs and data required by the arbitrary data processing apparatus, and may also be used for temporarily storing data that has been output or is to be output.
Example 4
The present embodiment relates to a computer program product for implementing the method of embodiment 1, comprising a computer program which, when executed by a processor, implements a CT image based soybean stalk internal structure identification method of embodiment 1.
Example 5
Referring to fig. 6, a soybean stalk internal structure recognition system based on CT images includes:
the soybean stalk CT image data acquisition module is used for placing soybean pot plants into plant CT acquisition equipment, and scanning soybean plants from top to bottom by utilizing micro-CT equipment to obtain acquired soybean stalk CT image data;
the image enhancement module is used for enhancing the acquired image in the common scanning mode, introducing the diffusion model into the image super-resolution network, inputting the acquired image into the CT image super-resolution network based on the diffusion model for enhancing, and obtaining a high-resolution tissue image;
the plant area image acquisition module is used for carrying out data augmentation processing on the processed enhanced image, inputting the enhanced image into the image segmentation network for plant area segmentation to obtain a mask image, and carrying out pixel point multiplication on the mask image and the corresponding CT image to obtain an image of a plant area;
and the plant internal phenotype data acquisition module is used for reconstructing the volume of the segmented CT image to obtain gridding data and accurately acquiring plant internal phenotype data.
Example 6
The present embodiment relates to a computing device for implementing the method of embodiment 1, including a memory and a processor, wherein the memory stores executable codes, and the processor implements the soybean stalk internal structure identification method based on CT images of embodiment 1 when executing the executable codes.
At the hardware level, the computing device includes a processor, internal bus, network interface, memory, and non-volatile storage, although other hardware required for the business is possible. The processor reads the corresponding computer program from the nonvolatile memory into the memory and then runs to implement the method described in embodiment 1 above. Of course, other implementations, such as logic devices or combinations of hardware and software, are not excluded from the present invention, that is, the execution subject of the following processing flows is not limited to each logic unit, but may be hardware or logic devices.
Improvements to one technology can clearly distinguish between improvements in hardware (e.g., improvements to circuit structures such as diodes, transistors, switches, etc.) and software (improvements to the process flow). However, with the development of technology, many improvements of the current method flows can be regarded as direct improvements of hardware circuit structures. Designers almost always obtain corresponding hardware circuit structures by programming improved method flows into hardware circuits. Therefore, an improvement of a method flow cannot be said to be realized by a hardware entity module. For example, a programmable logic device (Programmable Logic Device, PLD) (e.g., field programmable gate array (Field Programmable Gate Array, FPGA)) is an integrated circuit whose logic function is determined by the programming of the device by a user. A designer programs to "integrate" a digital system onto a PLD without requiring the chip manufacturer to design and fabricate application-specific integrated circuit chips. Moreover, nowadays, instead of manually manufacturing integrated circuit chips, such programming is mostly implemented by using "logic compiler" software, which is similar to the software compiler used in program development and writing, and the original code before the compiling is also written in a specific programming language, which is called hardware description language (Hardware Description Language, HDL), but not just one of the hdds, but a plurality of kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), lava, lola, myHDL, PALASM, RHDL (Ruby Hardware Description Language), etc., VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog are currently most commonly used. It will also be apparent to those skilled in the art that a hardware circuit implementing the logic method flow can be readily obtained by merely slightly programming the method flow into an integrated circuit using several of the hardware description languages described above.
The controller may be implemented in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer readable medium storing computer readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable logic controllers, and embedded microcontrollers, examples of which include, but are not limited to, the following microcontrollers: ARC 625D, atmel AT91SAM, microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic of the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller in a pure computer readable program code, it is well possible to implement the same functionality by logically programming the method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers, etc. Such a controller may thus be regarded as a kind of hardware component, and means for performing various functions included therein may also be regarded as structures within the hardware component. Or even means for achieving the various functions may be regarded as either software modules implementing the methods or structures within hardware components.
The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function. One typical implementation is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.
For convenience of description, the above devices are described as being functionally divided into various units, respectively. Of course, the functions of each element may be implemented in the same piece or pieces of software and/or hardware when implementing the present invention.
It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the foregoing detailed description of the invention has been provided, it will be apparent to those skilled in the art that modifications may be made to the embodiments described in the foregoing examples, and that certain features may be substituted for those illustrated and described herein. Modifications, equivalents, and alternatives falling within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (3)
1. The soybean stalk internal structure identification method based on the CT image is characterized by comprising the following steps of:
(1) Placing the soybean pot in a plant CT acquisition device, and scanning soybean plants from top to bottom by utilizing micro-CT equipment to obtain acquired soybean stalk CT image data;
(2) The method comprises the steps of performing enhancement processing on a picture acquired in a common scanning mode, introducing a diffusion model into an image super-resolution network, inputting the acquired picture into a CT image super-resolution network based on the diffusion model for enhancement processing, and obtaining a high-resolution tissue image;
(3) Performing data augmentation treatment on the treated enhanced image, inputting the enhanced image into an image segmentation network for plant region segmentation to obtain a mask image, and performing pixel point multiplication on the mask image and a corresponding CT image to obtain an image of a plant region;
(4) Performing volume reconstruction on the segmented CT image to obtain gridding data, and accurately obtaining plant internal phenotype data;
step (1), specifically comprising:
(1.1) collecting training data; the method comprises the steps of collecting plants by adopting common magnification and high-precision mode magnification respectively, and processing common magnification pictures and high-precision mode magnification pictures to obtain corresponding stalk structures, wherein the corresponding stalk structures are used as high-low resolution image pairs of the stalk structures;
(1.2) collecting test data; scanning plants to obtain a large-area acquisition image with lower resolution;
step (2), specifically comprising the following substeps:
(2.1) De-basedPerforming super-resolution network training by using the noise diffusion model; the low resolution CT input picture is LR, HR is the corresponding original high resolution CT picture, and the LR is subjected to a pre-training encoder model to obtain an initial prediction high resolution pictureInputting into a Unet network; UP (LR) is a picture directly after bicubic upsampling for low resolution, data initial state during diffusion +.>The difference value between the HR image and the UP-sampling picture UP (LR) contains high-frequency information; the diffusion process is directly based on->To arbitrary->Step->Sampling is carried out by the nature of the diffusion process, i.e. the sampling time step is +.>Variance of time->Actual sampling noise->Directly obtain->Of (1), wherein->Belonging to uniform distribution, i.e.，/>WhereinWherein->The variance used for each step is between 0 and 1; during training, will ∈ ->，/>And->Input the same to the Unet network to obtain estimated noise +.>And is +.>Comparing, calculating the loss by estimating the noise +.>Training and optimizing the noise estimation network, minimizing a loss function, and finally obtaining estimated noise (I) meeting convergence conditions>；
(2.2) performing super-resolution inference on the low-resolution image to obtain a high-resolution picture; each step ofCan only be obtained by the previous step; hidden variants obtained by passing LR images through an EncoderQuantity->Adding a conditional noise predictor to guide the generation of corresponding HR information while initially randomly sampling the noise figure +.>And->As input to the Unet, the calculation gets the +.>Noise ∈estimated by step>Then adding randomly sampled perturbations, obtaining +.>Is a noise figure of (1):whereinIn this way, circulation is continued until +.>Finally->Adding the UP-sampled image UP (LR) with LR to obtain a high-fraction CT picture SR;
step (3), specifically comprising:
(3.1) carrying out data augmentation processing on the processed high-resolution image, simulating different scanning parameters by using a contrast enhancement means, simulating different positioning postures during scanning by using random rotation, adding random up-down overturning and normalization operation, and normalizingThe chemical expression is:wherein->For inputting an image +.>For normalized image, ++>And->Maximum pixel value and minimum pixel value of the input image, respectively;
(3.2) obtaining different tissue divisions of the soybean stalk CT image; inputting the processed image into an open-source pre-training image segmentation network based on deep learning for training and optimizing so as to adapt to the CT image of the soybean stalk, and obtaining different tissue segmentations of the CT image of the soybean stalk through the segmentation network for training and optimizing;
step (4), specifically comprising:
(4.1) obtaining a continuous meshed three-dimensional curved surface; collecting the segmented CT image areas, obtaining three-dimensional data of soybean stalk tissues, and resampling and compressing the three-dimensional data by using a quadratic function sampling mode; extracting an isosurface by using a Marching cube algorithm, calculating a surface normal, acquiring triangular patches, and splicing the triangular patches to form a continuous grid three-dimensional curved surface;
(4.2) calculating a phenotypic parameter; and calculating thickness, included angle, surface area and volume parameters of the meshed three-dimensional curved surface, obtaining phenotype parameters of the partitioned areas, and accurately obtaining the phenotype data inside the plant.
2. A soybean stalk internal structure identification device based on CT images, comprising one or more processors configured to implement the soybean stalk internal structure identification method based on CT images of claim 1.
3. A computer-readable storage medium, having stored thereon a program which, when executed by a processor, implements the CT image-based soybean stalk internal structure recognition method of claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311278665.4A CN117011316B (en) | 2023-10-07 | 2023-10-07 | Method and system for identifying internal structure of soybean stalk based on CT image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311278665.4A CN117011316B (en) | 2023-10-07 | 2023-10-07 | Method and system for identifying internal structure of soybean stalk based on CT image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117011316A CN117011316A (en) | 2023-11-07 |
| CN117011316B true CN117011316B (en) | 2024-02-06 |
Family
ID=88571282
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311278665.4A Active CN117011316B (en) | 2023-10-07 | 2023-10-07 | Method and system for identifying internal structure of soybean stalk based on CT image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117011316B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022121160A1 (en) * | 2020-12-07 | 2022-06-16 | 苏州深透智能科技有限公司 | Method for enhancing quality and resolution of ct images based on deep learning |
| CN114677325A (en) * | 2022-01-25 | 2022-06-28 | 安徽农业大学 | Construction method of rice stem section segmentation model and detection method based on model |
| CN115019215A (en) * | 2022-08-09 | 2022-09-06 | 之江实验室 | Hyperspectral image-based soybean disease and pest identification method and device |
| CN116071239A (en) * | 2023-03-06 | 2023-05-05 | 之江实验室 | CT image super-resolution method and device based on mixed attention model |
| CN116485655A (en) * | 2023-05-10 | 2023-07-25 | 成都理工大学 | CT image super-resolution processing method based on diffusion model |
| CN116681584A (en) * | 2023-04-06 | 2023-09-01 | 长春理工大学 | Multistage diffusion image super-resolution algorithm |
-
2023
- 2023-10-07 CN CN202311278665.4A patent/CN117011316B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022121160A1 (en) * | 2020-12-07 | 2022-06-16 | 苏州深透智能科技有限公司 | Method for enhancing quality and resolution of ct images based on deep learning |
| CN114677325A (en) * | 2022-01-25 | 2022-06-28 | 安徽农业大学 | Construction method of rice stem section segmentation model and detection method based on model |
| CN115019215A (en) * | 2022-08-09 | 2022-09-06 | 之江实验室 | Hyperspectral image-based soybean disease and pest identification method and device |
| CN116071239A (en) * | 2023-03-06 | 2023-05-05 | 之江实验室 | CT image super-resolution method and device based on mixed attention model |
| CN116681584A (en) * | 2023-04-06 | 2023-09-01 | 长春理工大学 | Multistage diffusion image super-resolution algorithm |
| CN116485655A (en) * | 2023-05-10 | 2023-07-25 | 成都理工大学 | CT image super-resolution processing method based on diffusion model |
Non-Patent Citations (3)
| Title |
|---|
| Denoising Diffusion Probabilistic Models;Jonathan Ho 等;arXiv:2006.11239v2;全文 * |
| 基于CT图像的超分辨率重构研究;曹洪玉;刘冬梅;付秀华;张静;岳鹏飞;;长春理工大学学报(自然科学版)(01);全文 * |
| 基于植物根系 XCT 体数据的三维可视化重建;刘国成;南方职业教育学刊;第2-4页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117011316A (en) | 2023-11-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107784647B (en) | Liver and tumor segmentation method and system based on multitask deep convolutional network | |
| CN105957063B (en) | CT image liver segmentation method and system based on multiple dimensioned weighting similarity measure | |
| CN110047082B (en) | Deep learning-based pancreatic neuroendocrine tumor automatic segmentation method and system | |
| CN110276745B (en) | Pathological image detection algorithm based on generation countermeasure network | |
| CN111476793B (en) | Dynamic enhanced magnetic resonance imaging processing method, system, storage medium and terminal | |
| CN112258514B (en) | Segmentation method of pulmonary blood vessels of CT (computed tomography) image | |
| CN109242863B (en) | Ischemic stroke image region segmentation method and device | |
| CN110675411A (en) | Cervical squamous intraepithelial lesion recognition algorithm based on deep learning | |
| CN111260667B (en) | Neurofibromatosis segmentation method combined with spatial guidance | |
| CN112785603A (en) | Brain tissue segmentation method based on Unet and superpixel | |
| CN110570394A (en) | medical image segmentation method, device, equipment and storage medium | |
| CN112489062B (en) | Medical image segmentation method and system based on boundary and neighborhood guidance | |
| CN117011316B (en) | Method and system for identifying internal structure of soybean stalk based on CT image | |
| CN115867933A (en) | Computer-implemented method, computer program product and system for processing images | |
| CN111681168A (en) | Low-resolution cell super-resolution reconstruction method based on parallel residual error network | |
| CN116894783A (en) | Metal artifact removal method for countermeasure generation network model based on time-varying constraint | |
| CN116245892B (en) | Image processing model generation method, image processing method and device | |
| Rashid et al. | Single MR image super-resolution using generative adversarial network | |
| CN116091458A (en) | Pancreas image segmentation method based on complementary attention | |
| CN111292339B (en) | Clinical temporal bone CT multi-structure extraction method and device based on W-type network structure | |
| CN115661170A (en) | Method, device and medium for automatically segmenting abdomen three-dimensional CT image | |
| Du et al. | Expectation-maximization attention cross residual network for single image super-resolution | |
| CN115222651A (en) | Pulmonary nodule detection system based on improved Mask R-CNN | |
| CN116543246A (en) | Training method of image denoising model, image denoising method, device and equipment | |
| Chen et al. | Dynamic Degradation Intensity Estimation for Adaptive Blind Super-Resolution: A Novel Approach and Benchmark Dataset |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |