[go: up one dir, main page]

CN110400327B - A nighttime image segmentation method of tomato plants based on improved PCNN - Google Patents

A nighttime image segmentation method of tomato plants based on improved PCNN Download PDF

Info

Publication number
CN110400327B
CN110400327B CN201910663833.9A CN201910663833A CN110400327B CN 110400327 B CN110400327 B CN 110400327B CN 201910663833 A CN201910663833 A CN 201910663833A CN 110400327 B CN110400327 B CN 110400327B
Authority
CN
China
Prior art keywords
image
pcnn
image segmentation
link
tomato plant
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.)
Expired - Fee Related
Application number
CN201910663833.9A
Other languages
Chinese (zh)
Other versions
CN110400327A (en
Inventor
项荣
张杰兰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Jiliang University
Original Assignee
China Jiliang University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by China Jiliang University filed Critical China Jiliang University
Priority to CN201910663833.9A priority Critical patent/CN110400327B/en
Publication of CN110400327A publication Critical patent/CN110400327A/en
Application granted granted Critical
Publication of CN110400327B publication Critical patent/CN110400327B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

本发明公开了一种基于改进PCNN的番茄植株夜间图像分割方法。首先对采集的番茄植株彩色图像进行最大类间方差计算，并获取对应的最佳阈值level；基于最大类间方差的阈值改进PCNN算法，将最大类间方差的阈值直接赋值给PCNN模型中的链接权放大系数VE、阈值迭代衰减时间常数αE和突触间链接系数β；再将PCNN模型中的链接输入项进行改进，将神经元上一次的点火输出与内部连接矩阵经过卷积后再与最大类间方差的阈值进行相乘得到本次链接输入项。应用本发明可实现番茄植株夜间图像分割中PCNN模型参数的自适应调整，减少PCNN迭代次数，提高算法应用的实时性。

Figure 201910663833

The invention discloses a nighttime image segmentation method of tomato plants based on improved PCNN. First, calculate the maximum inter-class variance of the collected tomato plant color images, and obtain the corresponding optimal threshold level; improve the PCNN algorithm based on the threshold of the maximum inter-class variance, and directly assign the maximum inter-class variance threshold to the link in the PCNN model Weight amplification factor V E , threshold iterative decay time constant α E and intersynaptic link coefficient β; then the link input in the PCNN model is improved, and the last firing output of the neuron and the internal connection matrix are convolved and then Multiply with the threshold of the maximum inter-class variance to obtain the input item of this link. The application of the invention can realize the self-adaptive adjustment of the PCNN model parameters in the segmentation of the nighttime image of the tomato plant, reduce the number of iterations of the PCNN, and improve the real-time performance of the algorithm application.

Figure 201910663833

Description

Tomato plant night image segmentation method based on improved PCNN
Technical Field
The invention relates to a tomato plant night image segmentation method based on improved PCNN.
Background
Tomato production robot is the research focus in the current agricultural automation field, and as the mode that improves tomato production efficiency and realize tomato production automation, carrying out tomato production night is the effective way of the strong characteristics of tomato seasonality, and tomato production robot carries out work in daytime and night continuity and can be fine solution current tomato production in the labour resource shortage, the labour cost scheduling problem of being high. The identification and positioning of the tomato plant stems based on machine vision are the premise of picking tomatoes and trimming axillary buds and branches and leaves, and the realization of the image segmentation of the tomato plant stems is the premise of realizing the identification and positioning of the tomato plant stems. In order to improve the production efficiency of tomatoes, relevant production work is necessary at night, and the realization of the nighttime image segmentation of tomato plants is an important step for realizing the automation of a tomato production robot.
Under the illumination of the natural environment at night, the illumination is uneven, and shadows caused by branches and leaves are difficult to avoid, thereby bringing difficulty to the night image segmentation of the tomato plant stems.
The PCNN model is used for image processing, so that the neighborhood similar pixels in the target region and the background region can be well kept continuous. The problem of nighttime image segmentation of tomato plant stems is not solved by the conventional PCNN model, so that adjacent segmentation of organs of tomato plants is inaccurate, and the accuracy of identification of each organ of the tomato plants is influenced. In addition, the traditional PCNN model has the defects of high image segmentation iteration times and poor real-time performance, and is not enough to meet the requirements in production practice.
Disclosure of Invention
The invention aims to provide a tomato plant night image segmentation method based on improved PCNN, which reduces the iteration times in the traditional PCNN model image segmentation process, effectively reduces the image segmentation time, can realize the correct rate of tomato plant stem identification, and lays a foundation for identifying and positioning the tomato plant stems at night.
The technical scheme adopted by the invention is as follows:
illuminating nighttime tomato plants with an illumination system; the binocular stereo camera collects color images and transmits the images to image processing software for image segmentation; collecting a color image C of a tomato plant in a natural environment at night by using a binocular stereo camera; the method comprises the following steps:
preprocessing an image: carrying out noise reduction treatment on the tomato plant night color image C by using wiener filtering to obtain an image D;
improving the PCNN model: weighting the link input items in the PCNN model to obtain the improved PCNN model as shown in the formulas (1) to (5):
Fij(n)＝Sij (1)
Lij(n)＝t(∑klWijklYkl(n-1)) (2)
Uij(n)＝Fij(n)(1+βLij(n)) (3)
Figure GDA0002961305540000021
Figure GDA0002961305540000022
in the formula: fij-feedback input to the neuron; sij-grey scale image S corresponds to a pixel value; l isij-a link entry; u shapeij-an internal activity item; β -inter-synaptic linkage coefficient; y isij-a pulse output value; eij-a dynamic threshold; n-the number of iterations; t is a weight coefficient; i. j-represents a pixel in the digital image; k. l-neighborhood pixels representing a center pixel; alpha is alphaE-an iterative decay time constant of the dynamic threshold system; vE-a corresponding link weight amplification factor; wijklDetermined by the W link weight matrix, as shown in equation (6):
Figure GDA0002961305540000023
setting an initial value of the PCNN parameter, setting the iteration number n as 1, and starting iterative image segmentation based on the improved PCNN model;
fourthly, carrying out image segmentation on the image D subjected to noise reduction by using the improved PCNN model to obtain a binary image B (n);
calculating the information entropy P (n) of the binary image B (n);
sixthly, judging whether the information entropy P (n) is larger than the information entropy P (n-1) obtained in the previous iteration: if so, the iteration times are increased by 1, and the step IV is skipped; and otherwise, taking the image segmentation result B (n-1) obtained in the previous iteration as a final image segmentation result, and finishing the iterative image segmentation based on the improved PCNN model.
The tomato plant night image segmentation method based on the improved PCNN is characterized in that the weighting method for the chaining input items in the PCNN model in the step II comprises the following steps: obtaining an optimal threshold level for image segmentation of a tomato plant night color image C by using an OTSU image segmentation algorithm, taking the level value as a weight coefficient of a link input item, wherein the link input item in the weighted PCNN model is as shown in formula (7):
Lij(n)＝level(∑klWijklYkl(n-1)) (7)
in the formula: level-OTSU partition optimal threshold;
the tomato plant night image segmentation method based on the improved PCNN is characterized in that the PCNN parameter initial value setting method in the third step is as follows: obtaining an image segmentation threshold level of a tomato plant night color image C by using an OTSU image segmentation algorithm, and taking the level value as an iterative attenuation time constant alpha of a dynamic threshold systemEThe chain weight amplification factor VEAnd an initial value of the inter-synaptic linkage coefficient beta is shown in formula (8):
Figure GDA0002961305540000031
the invention has the beneficial effects that:
the method adopts the OTSU algorithm threshold value to assign values to the parameters in the PCNN model, realizes the self-adaptation of the parameters of the PCNN model, and adds the OTSU threshold value as a weight value in a link input item, so that the PCNN model only needs two iterations to obtain an ideal segmentation image, and the real-time performance of the PCNN model is greatly improved.
Drawings
Fig. 1 is a schematic diagram of a tomato plant night image acquisition and processing system.
FIG. 2 is a flow chart of algorithm segmentation of a tomato plant nighttime image improvement PCNN model.
Fig. 3 is a diagram of a modified PCNN model.
Fig. 4 is a schematic diagram of the link weight matrix W.
FIG. 5 is an improved PCNN model iterative segmentation graph.
Figure 6 is an OTSU algorithm segmentation graph.
In fig. 1: 1. tomato plant, 2 binocular stereo camera, 3, lighting system, 4, 1394 image acquisition card, 5, computer, 6, image processing software.
Detailed Description
The invention is further illustrated by the following figures and examples.
Fig. 1 illustrates a specific example of the illumination system for night image acquisition of tomato plants. The lighting system 3 employs an up-down lighting system consisting of 2 28w halogen lamps, with an up-down distance of 400 mm. The image receiving device adopts a binocular stereo camera 2 (the stereo camera can acquire three-dimensional position information of a target and is considered for acquiring the three-dimensional position information of a tomato plant component organ later), an image sensor in the binocular stereo camera 2 is a color Sony ICX204 CCD, the maximum resolution is 1024 multiplied by 768, and the focal length of a lens is 6 mm. The image acquisition card 4 is MOGE 1394 with a power adapter. The computer 5 is a DELL E4300 notebook computer with a memory of 2G, and the CPU is an Intel Core2 Duo P9400 WIN 7 operating system. The binocular stereo camera 2 is connected with a 1394 image acquisition card 4 by using a 1394 connecting wire, and the 1394 image acquisition card 4 is installed on a computer 5 through a 7-in-1 card reader interface.
The specific implementation of the PCNN model image segmentation algorithm for improving nighttime tomato plants is as follows:
illuminating the nighttime outdoor tomato plants 1 with an illumination system 3; a color CCD in the binocular stereo camera 2 receives a pair of optical image pairs of the tomato plant 1 and converts the optical image pairs into a pair of electronic image pairs for output; the pair of electronic images output by the binocular stereo camera 2 is input into a 1394 image acquisition card 4; the 1394 image acquisition card 4 converts the analog image signal into a digital image signal and inputs the digital image signal into the computer 5; the image processing software 6 in the computer 5 realizes the image segmentation processing on the tomato plants in the natural environment at night.
As shown in fig. 2, the improved PCNN model algorithm segmentation of the nighttime images of tomato plants is specifically realized as follows:
preprocessing an image: carrying out noise reduction treatment on the tomato plant night color image C by using wiener filtering to obtain an image D, wherein the wiener filtering is a self-adaptive linear filter with the least square as an optimal criterion; the filtering effect can be adjusted according to the local variance, the effect of removing Gaussian noise is obvious, and the method is suitable for removing noise of the tomato plant nighttime image;
improving the PCNN model: weighting the link input items in the PCNN model to obtain the improved PCNN model as shown in the formulas (1) to (5):
Fij(n)＝Sij (1)
Lij(n)＝t(∑klWijklYkl(n-1)) (2)
Uij(n)＝Fij(n)(1+βLij(n)) (3)
Figure GDA0002961305540000041
Figure GDA0002961305540000042
in the formula: fij-feedback input to the neuron; sij-grey scale image S corresponds to a pixel value; l isij-a link entry; u shapeij-an internal activity item; β -inter-synaptic linkage coefficient; y isij-a pulse output value; eij-a dynamic threshold; n-the number of iterations; t is a weight coefficient; i. j-represents a pixel in the digital image; k. l-neighborhood pixels representing a center pixel; alpha is alphaE-an iterative decay time constant of the dynamic threshold system; vE-a corresponding link weight amplification factor; wijklDetermined by the W link weight matrix, as shown in equation (6):
Figure GDA0002961305540000051
setting an initial value of the PCNN parameter, setting the iteration number n as 1, and starting iterative image segmentation based on the improved PCNN model;
fourthly, carrying out image segmentation on the image D subjected to noise reduction by using the improved PCNN model to obtain a binary image B (n);
calculating the information entropy P (n) of the binary image B (n); generally, the larger the maximum entropy of the image after segmentation, the larger the amount of information obtained from the original image after segmentation, the more detailed the segmented image, and therefore, the better the segmentation effect. As shown in formula (7):
H1(P)＝-P1log2P1-P0log2P0 (7)
sixthly, judging whether the information entropy P (n) is larger than the information entropy P (n-1) obtained in the previous iteration: if so, the iteration times are increased by 1, and the step IV is skipped; and otherwise, taking the image segmentation result B (n-1) obtained in the previous iteration as a final image segmentation result, and finishing the iterative image segmentation based on the improved PCNN model.
The tomato plant night image segmentation method based on the improved PCNN is characterized in that the weighting method for the chaining input items in the PCNN model in the step II comprises the following steps: obtaining an image segmentation threshold level of a tomato plant night color image C by using an OTSU image segmentation algorithm, wherein the formula (8) is as follows:
g＝max{ω0(μ0-μ)2+ω1(μ1-μ)2} (8)
in the formula: omega0-the number of pixels of the foreground in proportion to the whole image; omega1-the number of pixels of the background is proportional to the whole image; mu.s0-average gray level of foreground pixels; mu.s1-average gray level of background pixels; μ — total average gray scale of image pixels; g-maximum between-class variance; max-represents the maximum value;
the level is an optimal threshold corresponding to the maximum inter-class variance g of the image, the level value is used as a weight coefficient of a link input item, and the link input item in the weighted PCNN model is as shown in the formula (9):
Lij(n)＝level(∑klWijklYkl(n-1)) (9)
in the formula: level-OTSU splits the optimal threshold.
The tomato plant night image segmentation method based on the improved PCNN is characterized in that the PCNN parameter initial value setting method in the third step is as follows: obtaining an optimal threshold level of image segmentation of a tomato plant night color image C by using an OTSU image segmentation algorithm, and taking the level value as an iterative attenuation time constant alpha of a dynamic threshold systemEThe chain weight amplification factor VEInitial value of inter-synaptic linkage coefficient betaAs shown in formula (10):
Figure GDA0002961305540000061
through tests, the image segmentation result of the upper right corner of the image in the figure 5 is obtained based on a Matlab R2016a programming environment by applying the method, the total operation time of the algorithm is 1.126 seconds, and the maximum entropy value reaches 0.9309. In the improved PCNN model, the result of the first iteration is shown in the upper left corner of FIG. 5, and the maximum entropy value after segmentation is 0.0069; FIG. 5 shows the second iteration result in the upper right corner, after segmentation, the maximum entropy value is 0.9309, and the tomato, the stem and the leaves on the tomato plant are completely segmented; the lower left corner of fig. 5 is the result of the third iteration, the maximum entropy value after segmentation is 0.8026, and compared with the upper right corner of fig. 5, tomatoes and stalks in the lower left corner are all deleted after segmentation; the lower right corner of fig. 5 is the fourth iteration result, the maximum entropy value after segmentation is 0.7494, the tomato plant organ is seriously lost, and the noise in the image is serious; because the PCNN iterative segmentation image quality is from poor to excellent and from good to poor, the second iterative effect in the improved PCNN model is the best, and the second iterative result is selected as the image segmentation result of the improved PCNN model. FIG. 6 shows the result of OTSU segmentation, which has a maximum entropy value of 0.7151.

Claims (3)

1. A tomato plant night image segmentation method based on improved PCNN is characterized by comprising the following steps:
preprocessing an image: carrying out noise reduction treatment on the tomato plant night color image C by using wiener filtering to obtain an image D;
improving the PCNN model: weighting the link input items in the PCNN model to obtain the improved PCNN model as shown in the formulas (1) to (5):
Fij(n)＝Sij (1)
Lij(n)＝t(∑klWijklYkl(n-1)) (2)
Uij(n)＝Fij(n)(1+βLij(n)) (3)
Figure FDA0002961305530000011
Figure FDA0002961305530000012
in the formula: fij-feedback inputs of neurons; sij-the grey image S corresponds to pixel values; l isij-a link entry; u shapeij-an internal activity item; β -inter-synaptic linkage coefficient; y isij-a pulse output value; eij-a dynamic threshold; n-number of iterations; t-weight coefficient; i. j-represents a certain pixel in the digital image; k. l-neighborhood pixels representing a center pixel; alpha is alphaE-an iterative decay time constant of the dynamic threshold system; vE-a corresponding link weight amplification factor; wijklDetermined by the W link weight matrix, as shown in equation (6):
Figure FDA0002961305530000013
setting an initial value of the PCNN parameter, setting the iteration number n as 1, and starting iterative image segmentation based on the improved PCNN model;
fourthly, carrying out image segmentation on the image D subjected to noise reduction by using the improved PCNN model to obtain a binary image B (n);
calculating the information entropy P (n) of the binary image B (n);
sixthly, judging whether the information entropy P (n) is larger than the information entropy P (n-1) obtained in the previous iteration: if so, the iteration times are increased by 1, and the step IV is skipped; and otherwise, taking the image segmentation result B (n-1) obtained in the previous iteration as a final image segmentation result, and finishing the iterative image segmentation based on the improved PCNN model.
2. The improved PCNN-based tomato plant nighttime image segmentation method of claim 1, wherein the weighting method for the link entries in the PCNN model comprises: obtaining an optimal threshold level for image segmentation of a tomato plant night color image C by using an OTSU image segmentation algorithm, taking the level value as a weight coefficient of a link input item, wherein the link input item in the weighted PCNN model is as shown in formula (7):
Lij(n)＝level(∑klWijklYkl(n-1)) (7)
in the formula: level-OTSU partitions the optimal threshold.
3. The improved PCNN-based tomato plant night image segmentation method as claimed in claim 1, wherein the PCNN parameter initial value setting method in step (iii) comprises: obtaining an optimal threshold level of image segmentation of a tomato plant night color image C by using an OTSU image segmentation algorithm, and taking the level value as an iterative attenuation time constant alpha of a dynamic threshold systemEThe chain weight amplification factor VEAnd an initial value of the inter-synaptic linkage coefficient beta is shown in formula (8):
Figure FDA0002961305530000021
CN201910663833.9A 2019-07-23 2019-07-23 A nighttime image segmentation method of tomato plants based on improved PCNN Expired - Fee Related CN110400327B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910663833.9A CN110400327B (en) 2019-07-23 2019-07-23 A nighttime image segmentation method of tomato plants based on improved PCNN

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910663833.9A CN110400327B (en) 2019-07-23 2019-07-23 A nighttime image segmentation method of tomato plants based on improved PCNN

Publications (2)

Publication Number Publication Date
CN110400327A CN110400327A (en) 2019-11-01
CN110400327B true CN110400327B (en) 2021-07-02

Family

ID=68325252

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910663833.9A Expired - Fee Related CN110400327B (en) 2019-07-23 2019-07-23 A nighttime image segmentation method of tomato plants based on improved PCNN

Country Status (1)

Country Link
CN (1) CN110400327B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2004097535A (en) * 2002-09-10 2004-04-02 Toshiba Corp Method for segmenting medical three-dimensional image data
CN102999905A (en) * 2012-11-15 2013-03-27 天津工业大学 Automatic eye fundus image vessel detecting method based on PCNN (pulse coupled neural network)
CN107154048A (en) * 2017-05-09 2017-09-12 中国科学院遥感与数字地球研究所 The remote sensing image segmentation method and device of a kind of Pulse-coupled Neural Network Model
CN107274425A (en) * 2017-05-27 2017-10-20 三峡大学 A kind of color image segmentation method and device based on Pulse Coupled Neural Network
CN108182685A (en) * 2017-12-27 2018-06-19 三峡大学 The image partition method and device of a kind of combination belief propagation thought and PCNN

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4208485B2 (en) * 2001-05-31 2009-01-14 キヤノン株式会社 Pulse signal processing circuit, parallel processing circuit, pattern recognition device, and image input device

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2004097535A (en) * 2002-09-10 2004-04-02 Toshiba Corp Method for segmenting medical three-dimensional image data
CN102999905A (en) * 2012-11-15 2013-03-27 天津工业大学 Automatic eye fundus image vessel detecting method based on PCNN (pulse coupled neural network)
CN107154048A (en) * 2017-05-09 2017-09-12 中国科学院遥感与数字地球研究所 The remote sensing image segmentation method and device of a kind of Pulse-coupled Neural Network Model
CN107274425A (en) * 2017-05-27 2017-10-20 三峡大学 A kind of color image segmentation method and device based on Pulse Coupled Neural Network
CN108182685A (en) * 2017-12-27 2018-06-19 三峡大学 The image partition method and device of a kind of combination belief propagation thought and PCNN

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
DENG Xiangyu et al. .PCNN model analysis and its automatic parameters determination in image segmentation and edge detection.《Chinese Journal of Electronics》.2014,97-103. *
Image segmentation for whole tomato plant recognition at night;RongXiang et al.;《Computers and Electronics in Agriculture》;20181130;434-442 *
一种基于PCNN和改进的OTSU的图像分割算法;张松 等;《中国科技论文》;20160131;第11卷(第2期);236-240 *
一种复杂背景环境下的改进型PCNN图像分割算法;刘军 等;《计算机与数字工程》;20180220;375-406 *

Also Published As

Publication number Publication date
CN110400327A (en) 2019-11-01

Similar Documents

Publication Publication Date Title
CN109308697B (en) Leaf disease identification method based on machine learning algorithm
CN109360206B (en) Method for dividing rice ears in field based on deep learning
CN113919443B (en) Tobacco maturity state probability calculation method based on image analysis
CN109086823B (en) Automatic statistical method for wheat scab ear disease rate
CN114140692B (en) Fresh corn maturity prediction method based on UAV remote sensing and deep learning
CN119580049B (en) System for realizing intelligent recognition of crops by multi-mode method based on CLIP
CN116703932A (en) CBAM-HRNet model wheat spike grain segmentation and counting method based on convolution attention mechanism
CN113298777A (en) Cotton leaf blight detection method and system based on color features and super-pixel clustering
CN116524279A (en) Artificial intelligent image recognition crop growth condition analysis method for digital agriculture
CN119399635A (en) A soybean leaf disease recognition method based on multi-scale feature fusion Transformer
CN118397030B (en) Machine vision-based vegetable crop real-time water consumption online monitoring system and method
CN117831074A (en) A method for detecting dead chickens in cages based on deep learning
Jin et al. An improved mask R-CNN method for weed segmentation
CN118864305A (en) Animal body temperature recognition system for animal husbandry and veterinary medicine
CN118298467A (en) Intelligent recognition method and system for individual anomalies of breeding hens based on machine vision
CN108090910B (en) A Nighttime Outdoor Tomato Plant Image Segmentation Algorithm Based on Information Entropy Gradient Simplified PCNN Model
CN110400327B (en) A nighttime image segmentation method of tomato plants based on improved PCNN
CN105118028A (en) Automatic grafting machine grafted seedling seam identifying apparatus and seam identifying method
CN111967357A (en) Intelligent sorghum disease identification system and identification method based on machine vision
CN116071268A (en) Image Deillumination Model Based on Contrastive Learning and Its Training Method
CN119927935B (en) A vibration frequency control method and device for a vibration-type oil-tea picking robot
CN120495661A (en) Tomato plant segmentation method based on improved yolov s-seg model
CN117853738B (en) Image processing method and device for tea grading
CN119600400A (en) Detection method and detection device for diseased fish
CN109712110B (en) Crop real-time water consumption monitoring method and device based on machine vision

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
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20210702