CN113255434B - Apple identification method integrating fruit characteristics and deep convolutional neural network - Google Patents
Apple identification method integrating fruit characteristics and deep convolutional neural network Download PDFInfo
- Publication number
- CN113255434B CN113255434B CN202110376545.2A CN202110376545A CN113255434B CN 113255434 B CN113255434 B CN 113255434B CN 202110376545 A CN202110376545 A CN 202110376545A CN 113255434 B CN113255434 B CN 113255434B
- Authority
- CN
- China
- Prior art keywords
- fruit
- shape
- adopting
- frame
- image
- 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
- 235000013399 edible fruits Nutrition 0.000 title claims abstract description 80
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 15
- 241000220225 Malus Species 0.000 claims abstract description 52
- 235000021016 apples Nutrition 0.000 claims abstract description 11
- 238000010586 diagram Methods 0.000 claims description 19
- 238000001514 detection method Methods 0.000 claims description 9
- 238000009826 distribution Methods 0.000 claims description 9
- 230000003595 spectral effect Effects 0.000 claims description 6
- 230000010354 integration Effects 0.000 claims description 5
- 238000001228 spectrum Methods 0.000 claims description 5
- 238000003709 image segmentation Methods 0.000 claims description 3
- 238000003064 k means clustering Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000002156 mixing Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 description 3
- 238000012271 agricultural production Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000000746 purification Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000004345 fruit ripening Effects 0.000 description 1
- 235000012055 fruits and vegetables Nutrition 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000002420 orchard Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention discloses an apple identification method integrating fruit features and a deep convolutional neural network, which takes a Faster RCNN as a basic network, and integrates the fruit features into an input layer, an RPN and a position regression branch of the Faster RCNN, so that the customization of an apple identification framework is realized, and the accuracy of apple identification and positioning is effectively improved. The algorithm fully considers the color, shape and other characteristics of apples, and plays an important role in improving the fruit recognition accuracy under the conditions of adhesion and shielding.
Description
Technical Field
The invention relates to the field of computer vision and agricultural engineering, in particular to an apple identification method used in natural environment, and specifically relates to an apple identification method integrating fruit characteristics and a deep convolutional neural network.
Background
A large amount of labor force is needed to carry out concentrated picking operation in the fruit ripening period, so that the labor force is required to be large, and meanwhile, the time is required to be concentrated. However, the population of agriculture in China is seriously aged at present, and the population actually engaged in agricultural production is drastically reduced and the labor cost of agricultural production is remarkably improved in fifteen years. Therefore, it is urgent to develop a high-efficiency picking robot to replace manual picking operation of fruits and vegetables.
The identification and positioning of apples is the focus of research on apple picking robots. The fruit recognition method based on the deep convolutional neural network is a mainstream method for carrying out fruit recognition at present, and the speed and effect of recognition are superior to those of the traditional fruit recognition method. However, most fruit identification methods based on deep convolutional neural networks do not comprehensively consider the characteristics of fruits, and the accuracy of identification and positioning is difficult to further improve. Because the orchard environment is complex, the fruit is influenced by various factors such as light, adhesion and shielding, the structure of the deep convolutional neural network needs to be improved and optimized by combining the fruit characteristics, and therefore the accuracy of recognition and positioning is effectively improved.
Disclosure of Invention
Aiming at the technical problems, the technical scheme provides the apple identification method for fusing the fruit characteristics and the deep convolutional neural network, which can effectively solve the problems.
The invention is realized by the following technical scheme:
the apple identification method integrating fruit features and a deep convolutional neural network takes a Faster RCNN as a basic network, and the apple identification framework is customized by integrating the fruit features into an input layer, an RPN and a position regression branch of the Faster RCNN, and the apple identification framework comprises the following steps:
step 1: taking a Faster RCNN target detection frame as a basic network frame, and blending fruit characteristics into an input layer, RPN and position regression branches of the Faster RCNN;
step 2: the method comprises the steps of expressing the color, shape, texture, edge, spectrum and three-dimensional characteristics of apples in a graphical mode, and integrating the color, shape, texture, edge, spectrum and three-dimensional characteristics with fruit images to form a multi-channel image as input of a target detection frame;
step 3: determining a possible area of the fruit according to the color characteristics of the fruit, and limiting the generation position of a pre-selection frame in the RPN to the area;
step 4: the shape of a pre-selected frame in the RPN is set to be round, so that the shape of the pre-selected frame is matched with that of an apple, and therefore, the characteristics of purified fruits are effectively shielded from partial background interference;
step 5: setting the size of a preselected frame according to the size distribution of apple fruits in the image so as to be matched with the size of the fruits;
step 6: and replacing branches for rectangular frame position regression in the Faster RCNN framework by adopting a branch structure capable of carrying out ellipse shape prediction and position regression.
Further, the fruit features in the step 2 are graphically expressed, wherein color features are expressed by adopting a color difference map or a cluster map, shape features are expressed by adopting a gradient map, texture features are expressed by adopting a coding map, edge features are expressed by adopting a side map, spectral features are expressed by adopting a multi-layer spectral reflection map, and three-dimensional shape features are expressed by adopting a point cloud map or an intensity map.
Further, the integration of the fruit image and the multichannel image described in step 2 is used as the input of the target detection frame, and the specific integration method is as follows: superimposing the patterned features of the fruit onto a third dimension of the fruit image to form an image containing a plurality of channels, and modifying the number of channels of the Faster RCNN input layer according to the number of channels of the image.
Further, the determining the possible existence area of the fruit according to the color characteristics of the fruit in the step 3 comprises the following specific operation steps:
firstly, generating a color difference map Gray of an apple image by adopting an R-G operator, wherein the specific R-G operator is as follows:
then, 1/10 of the obtained optimal threshold value is calculated by adopting a maximum inter-class variance method to carry out image segmentation, and the connected region is expanded to obtain a region where apples possibly exist; the maximum inter-class variance method classifies pixels in an image into 2 classes by finding an optimal threshold value Thr, and the specific formula is as follows:
wherein the method comprises the steps ofAnd->Inter-class variance and intra-class variance of the 2-class pixel values, respectively; mu (mu) 1 Sum mu 2 Respectively the average value of the pixel values of class 2; p is p 1 And p 2 The proportion of the number of the class 2 pixels to the total number of the pixels is respectively; />And->Respectively the variance of the class 2 pixel values.
Further, the shape of the pre-selected frame in the RPN is set to be circular in the step 4, and the specific operation steps are as follows:
firstly, the shape of a pre-selected frame in the RPN is set to be square and used for participating in operation, and the specific operation mode is as follows: generating 3 square pre-selected frames with different sizes by taking each pixel of a possible fruit existence area as a center, and then mapping the square pre-selected frames onto a feature map according to the size proportion of an input image and the feature map;
and when the features in the preselected frame are extracted, the shape of the preselected frame is modified into a round shape by adopting a mask pattern, so that the shape of the preselected frame is matched with the shape of an apple.
Further, setting the size of a preselected frame in the step 5, firstly, marking the size of fruits in an image on a large scale, then analyzing and drawing a size distribution diagram of the fruits of the sample by adopting a clustering algorithm, and finally, selecting a proper value in the diagram as the size of the preselected frame; the specific clustering algorithm is as follows: first in the size distribution range x 1 ,x 2 ]Inner flat3 seed points are selected, and the values of the seed points are respectively x1+ h, x1+3h and x1+5h, wherein h= (x) 2 -x 1 ) 6; and then clustering the seeds by adopting a k-means clustering algorithm to determine the proper size of the preselected frame.
Further, the specific operation steps of the elliptic regression branch in the step 6 are as follows:
firstly, realizing pixel level segmentation of a local area based on a full convolution network, and then predicting the center coordinates of the ellipse, the radius of a long shaft, the inclination angle and the offset through an inserted parameter regression module.
(III) beneficial effects
Compared with the prior art, the apple identification method integrating fruit characteristics and the deep convolutional neural network has the following beneficial effects:
(1) Under the frame of the deep convolutional neural network, the technical scheme effectively fuses various characteristics such as color, texture and shape of the fruits, particularly designs a circular pre-selection frame and an elliptical prediction branch structure aiming at the shape characteristics of the fruits, improves the recognition and positioning accuracy of the fruits, and is more remarkable in adhesion and shielding of the fruits.
(2) According to the technical scheme, the input of the features is increased on the basis of the Faster RCNN, so that the features of the input network are more diversified, the convolutional network can learn the corresponding features, and the recognition accuracy is improved. Meanwhile, the RPN network is improved, the shape and the size of the preselection frame and the generation mode are mainly modified, so that the preselection frame is more matched with the shape of an apple, and the purification of target characteristics is facilitated. Besides, a branch capable of carrying out elliptic regression and position prediction is provided to replace the original rectangular positioning frame regression branch, so that the network can better identify and position adhered fruits and shade fruits.
Drawings
Fig. 1 is an overall structure diagram of a fruit recognition method according to the present invention.
FIG. 2 is a block diagram of a Faster RCNN target detection framework in the present invention.
Fig. 3a is a representation of the color profile of an apple image in accordance with the present invention.
Fig. 3b is a representation of the shape characteristics of an apple image in the present invention.
Fig. 3c is an edge feature representation of an apple image in accordance with the present invention.
Fig. 3d is a representation of the stereoscopic shape characteristics of an apple image in the present invention.
Fig. 4a is an original view of an apple in the present invention.
FIG. 4b is a graph of color differences represented by the R-G operator for apples in the present invention.
Fig. 4c is a diagram illustrating an exemplary area where apples may be present in the present invention.
Fig. 5 is a mask pattern illustration of an apple in the present invention.
Fig. 6 is a graph showing the recognition effect of apples in the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. The described embodiments are only some, but not all, embodiments of the invention. Various modifications and improvements of the technical scheme of the invention, which are made by those skilled in the art, are included in the protection scope of the invention without departing from the design concept of the invention.
Example 1:
the apple identification method integrating fruit features and a deep convolutional neural network takes a Faster RCNN as a basic network, and the apple identification framework is customized by integrating the fruit features into an input layer, an RPN and a position regression branch of the Faster RCNN, and the apple identification framework comprises the following steps:
step 1: taking a Faster RCNN target detection framework as a basic network framework, as shown in FIG. 2; on this basis, the fruit features are integrated into the input layer, RPN and position regression branches of the Faster RCNN, and the overall structure diagram of the Faster RCNN after the integration of the fruit features is shown in FIG. 1.
As can be seen by comparing fig. 1 and 2:
(1) The input of the features is increased on the basis of the Faster RCNN, so that the features of the input network are more diversified, the convolutional network can learn the corresponding features, and the recognition accuracy is improved.
(2) The RPN network is improved, the shape and the size of the preselection frame and the generation mode are mainly modified, so that the preselection frame is more matched with the shape of an apple, and the purification of target characteristics is facilitated.
(3) The branch capable of carrying out elliptic regression and position prediction is provided to replace the original rectangular positioning frame regression branch, so that the network can better identify and position adhesion and shade fruits.
Step 2: the color, shape, texture, edge, spectrum and stereo characteristics of apples are expressed in a graphical mode, and are integrated with fruit images to form a multi-channel image which is used as input of a target detection frame. The color features are expressed by adopting a color difference diagram or a cluster diagram, the shape features are expressed by adopting a gradient diagram, the texture features are expressed by adopting a coding diagram, the edge features are expressed by adopting a side diagram, the spectral features are expressed by adopting a multi-layer spectral reflection diagram, and the three-dimensional shape features are expressed by adopting a point cloud diagram or an intensity diagram.
The specific integration mode is as follows: superimposing the patterned features of the fruit onto a third dimension of the fruit image to form an image containing a plurality of channels, and modifying the number of channels of the Faster RCNN input layer according to the number of channels of the image.
Fig. 3a is a three-channel color apple image, fig. 3b adopts an R-G color operator to graphically express the color characteristics of the apple, fig. 3c adopts a canny operator to graphically express the edge characteristics of the fruit image, and fig. 3d adopts a depth map to graphically express the stereoscopic characteristics of the fruit image.
Step 3: the area where the fruit may exist is determined according to the color characteristics of the fruit, and the generation position of the pre-selection frame in the RPN is limited to the area. Firstly, generating a color difference map Gray of an apple image by adopting an R-G operator, wherein the specific R-G operator is as follows:
then, 1/10 of the obtained optimal threshold value is calculated by adopting a maximum inter-class variance method to carry out image segmentation, and the connected region is expanded to obtain a region where apples possibly exist; the maximum inter-class variance method classifies pixels in an image into 2 classes by finding an optimal threshold value Thr, and the specific formula is as follows:
wherein the method comprises the steps ofAnd->Inter-class variance and intra-class variance of the 2-class pixel values, respectively; mu (mu) 1 Sum mu 2 Respectively the average value of the pixel values of class 2; p is p 1 And p 2 The proportion of the number of the class 2 pixels to the total number of the pixels is respectively; />And->Respectively the variance of the class 2 pixel values.
The area accounts for about 1/5 of the total area of the image, and the generation of the pre-selection frame in the area can greatly reduce the calculation amount of the RPN network and reduce the risk of false identification; examples of the possible existence area of the fruit are shown in FIG. 4, FIG. 4b is a color difference diagram represented by R-G operator, and FIG. 4c is the possible existence area of the fruit.
Step 4: the shape of the pre-selected box in the RPN is set to be circular, so that the pre-selected box is matched with the shape of the apple, and therefore, the characteristics of the purified fruit are effectively shielded from partial background interference. Firstly, the shape of a pre-selected frame in the RPN is set to be square and used for participating in operation, and the specific operation mode is as follows: generating 3 square pre-selected frames with different sizes by taking each pixel of a possible fruit existence area as a center, and then mapping the square pre-selected frames onto a feature map according to the size proportion of an input image and the feature map;
and when the features in the preselected frame are extracted, the shape of the preselected frame is modified into a round shape by adopting a mask pattern, so that the shape of the preselected frame is matched with the shape of an apple. Fig. 5 is a mask pattern, which is overlaid on a corresponding square pre-selected frame, only the white areas in the mask pattern remain, and the black areas do not remain.
Step 5: according to the size distribution of the apple fruits in the image, the size of the preselected frame is set to be matched with the size of the fruits. Firstly, the sizes of fruits in the images are marked on a large scale, then, a clustering algorithm is adopted to analyze and draw a size distribution diagram of the fruits of the sample, and finally, a proper value is selected in the diagram to serve as a preselected frame size. The specific clustering algorithm is as follows:
first in the size distribution range x 1 ,x 2 ]3 seed points are selected on average, and the values of the seed points are respectively x1+ h, x1+3h and x1+5h, wherein h= (x) 2 -x 1 ) 6; and then clustering the seeds by adopting a k-means clustering algorithm to determine the proper size of the preselected frame.
Step 6: and replacing branches for rectangular frame position regression in the Faster RCNN framework by adopting a branch structure capable of carrying out ellipse shape prediction and position regression.
Firstly, realizing pixel level segmentation of a local area based on a full convolution network, and then predicting the center coordinates of the ellipse, the radius of a long shaft, the inclination angle and the offset through an inserted parameter regression module. An example of the final recognition effect is shown in fig. 6.
Claims (3)
1. An apple identification method integrating fruit characteristics and a deep convolutional neural network is characterized by comprising the following steps of: the method takes a Faster RCNN as a basic network, and the fruit features are fused into an input layer, an RPN and a position regression branch of the Faster RCNN, so that the customization of an apple identification framework is realized, and the method comprises the following steps:
step 1: taking a FasterRCNN target detection framework as a basic network framework, and blending fruit characteristics into an input layer, RPN and position regression branches of the FasterRCNN;
step 2: the method comprises the steps of expressing the color, shape, texture, edge, spectrum and three-dimensional characteristics of apples in a graphical mode, and integrating the color, shape, texture, edge, spectrum and three-dimensional characteristics with fruit images to form a multi-channel image as input of a target detection frame; wherein, the multi-channel image is integrated with the fruit image as the input of the target detection frame, and the specific integration mode is as follows: superimposing the graphical features of the fruit on a third dimension of the fruit image to form an image containing a plurality of channels, and modifying the number of channels of the Faster RCNN input layer according to the number of channels of the image;
step 3: determining a possible area of the fruit according to the color characteristics of the fruit, and limiting the generation position of a pre-selection frame in the RPN to the area; wherein, the possible existence area of the fruit is determined according to the color characteristics of the fruit, and the specific operation steps are as follows:
firstly, generating a color difference map Gray of an apple image by adopting an R-G operator, wherein the specific R-G operator is as follows:
then, 1/10 of the obtained optimal threshold value is calculated by adopting a maximum inter-class variance method to carry out image segmentation, and the connected region is expanded to obtain a region where apples possibly exist; the maximum inter-class variance method classifies pixels in an image into 2 classes by finding an optimal threshold value Thr, and the specific formula is as follows:
wherein,and->Inter-class variance and intra-class variance of the 2-class pixel values, respectively; mu (mu) 1 Sum mu 2 Respectively the average value of the pixel values of class 2; p is p 1 And p 2 The proportion of the number of the class 2 pixels to the total number of the pixels is respectively; />And->Variance of class 2 pixel values, respectively;
step 4: the shape of a pre-selected frame in the RPN is set to be round, so that the shape of the pre-selected frame is matched with that of an apple, and therefore, the characteristics of purified fruits are effectively shielded from partial background interference; wherein, the shape of the preselection frame in the RPN is set to be round, and the specific operation steps are as follows:
firstly, the shape of a pre-selected frame in the RPN is set to be square and used for participating in operation, and the specific operation mode is as follows: generating 3 square pre-selected frames with different sizes by taking each pixel of a possible fruit existence area as a center, and then mapping the square pre-selected frames onto a feature map according to the size proportion of an input image and the feature map;
when the features in the preselected frame are extracted, the shape of the preselected frame is modified into a round shape by adopting a mask graph, so that the shape of the preselected frame is matched with the shape of an apple;
step 5: setting the size of a preselected frame according to the size distribution of apple fruits in the image so as to be matched with the size of the fruits; wherein, the setting preselection frame size, specific operation mode is:
firstly, marking the sizes of fruits in an image on a large scale, then adopting a clustering algorithm to analyze and draw a size distribution diagram of the fruits of a sample, and finally selecting a proper value in the diagram as a preselected frame size; the specific clustering algorithm is as follows: first in the size distribution range x 1 ,x 2 ]3 seed points are selected on average, and the values of the seed points are respectively x1+ h, x1+3h and x1+5h, wherein h= (x) 2 -x 1 ) 6; then adopting a k-means clustering algorithm to perform clustering by taking a seed point as a center to determine a proper pre-selected frame size;
step 6: and replacing branches for rectangular frame position regression in the Faster RCNN framework by adopting a branch structure capable of carrying out ellipse shape prediction and position regression.
2. The apple identification method integrating fruit features with a deep convolutional neural network according to claim 1, wherein: and (2) graphically expressing the fruit characteristics, wherein color characteristics are expressed by adopting a color difference map or a cluster map, shape characteristics are expressed by adopting a gradient map, texture characteristics are expressed by adopting a coding map, edge characteristics are expressed by adopting a side map, spectral characteristics are expressed by adopting a multilayer spectral reflection map, and three-dimensional shape characteristics are expressed by adopting a point cloud map or an intensity map.
3. The apple identification method integrating fruit features with a deep convolutional neural network according to claim 1, wherein: and 6, carrying out elliptic regression branching, namely firstly realizing pixel level segmentation of a local area based on a full convolution network, and then predicting the center coordinates, the radius of a long shaft, the inclination angle and the offset of the ellipse through an inserted parameter regression module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110376545.2A CN113255434B (en) | 2021-04-08 | 2021-04-08 | Apple identification method integrating fruit characteristics and deep convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110376545.2A CN113255434B (en) | 2021-04-08 | 2021-04-08 | Apple identification method integrating fruit characteristics and deep convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113255434A CN113255434A (en) | 2021-08-13 |
| CN113255434B true CN113255434B (en) | 2023-12-19 |
Family
ID=77220466
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110376545.2A Active CN113255434B (en) | 2021-04-08 | 2021-04-08 | Apple identification method integrating fruit characteristics and deep convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113255434B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114612769B (en) * | 2022-03-14 | 2023-05-26 | 电子科技大学 | Integrated sensing infrared imaging ship detection method integrated with local structure information |
| CN114913179B (en) * | 2022-07-19 | 2022-10-21 | 南通海扬食品有限公司 | Apple skin defect detection system based on transfer learning |
| CN116021526A (en) * | 2023-02-07 | 2023-04-28 | 台州勃美科技有限公司 | Agricultural robot control method and device and agricultural robot |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105718945A (en) * | 2016-01-20 | 2016-06-29 | 江苏大学 | Apple picking robot night image identification method based on watershed and nerve network |
| CN108470339A (en) * | 2018-03-21 | 2018-08-31 | 华南理工大学 | A kind of visual identity of overlapping apple and localization method based on information fusion |
| WO2019040866A2 (en) * | 2017-08-25 | 2019-02-28 | The Board Of Trustees Of The University Of Illinois | Apparatus and method for agricultural data collection and agricultural operations |
| CN109948444A (en) * | 2019-02-19 | 2019-06-28 | 重庆理工大学 | Method for synchronously recognizing, system and the robot of fruit and barrier based on CNN |
| CN111325764A (en) * | 2020-02-11 | 2020-06-23 | 广西师范大学 | Fruit image contour recognition method |
| JP2020205048A (en) * | 2019-06-18 | 2020-12-24 | 富士通株式会社 | Object detection method based on deep learning network, apparatus, and electronic device |
| CN112395972A (en) * | 2020-11-16 | 2021-02-23 | 中国科学院沈阳自动化研究所 | Electric power system insulator string identification method based on unmanned aerial vehicle image processing |
-
2021
- 2021-04-08 CN CN202110376545.2A patent/CN113255434B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105718945A (en) * | 2016-01-20 | 2016-06-29 | 江苏大学 | Apple picking robot night image identification method based on watershed and nerve network |
| WO2019040866A2 (en) * | 2017-08-25 | 2019-02-28 | The Board Of Trustees Of The University Of Illinois | Apparatus and method for agricultural data collection and agricultural operations |
| CN108470339A (en) * | 2018-03-21 | 2018-08-31 | 华南理工大学 | A kind of visual identity of overlapping apple and localization method based on information fusion |
| CN109948444A (en) * | 2019-02-19 | 2019-06-28 | 重庆理工大学 | Method for synchronously recognizing, system and the robot of fruit and barrier based on CNN |
| JP2020205048A (en) * | 2019-06-18 | 2020-12-24 | 富士通株式会社 | Object detection method based on deep learning network, apparatus, and electronic device |
| CN111325764A (en) * | 2020-02-11 | 2020-06-23 | 广西师范大学 | Fruit image contour recognition method |
| CN112395972A (en) * | 2020-11-16 | 2021-02-23 | 中国科学院沈阳自动化研究所 | Electric power system insulator string identification method based on unmanned aerial vehicle image processing |
Non-Patent Citations (4)
| Title |
|---|
| "多功能果蔬采摘机器人的图像识别算法研究";刘晓洋;《中国博士学位论文全文数据库》;全文 * |
| "A Detection Method for Apple Fruits Based on Color and Shape Features";Xiaoyang Liu;《IEEE Access》;全文 * |
| "复杂环境下苹果采摘机器人的目标识别算法";田玉宇;《中国优秀硕士学位论文全文数据库》;全文 * |
| 基于颜色与果径特征的苹果树果实检测与分级;樊泽泽;柳倩;柴洁玮;杨晓峰;李海芳;;计算机工程与科学(09);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113255434A (en) | 2021-08-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113255434B (en) | Apple identification method integrating fruit characteristics and deep convolutional neural network | |
| CN105718945B (en) | Apple picking robot night image recognition method based on watershed and neural network | |
| Fu et al. | A novel image processing algorithm to separate linearly clustered kiwifruits | |
| CN108009542B (en) | Weed image segmentation method in rape field environment | |
| Xiang et al. | Recognition of clustered tomatoes based on binocular stereo vision | |
| Xiong et al. | The recognition of litchi clusters and the calculation of picking point in a nocturnal natural environment | |
| CN110232389B (en) | Stereoscopic vision navigation method based on invariance of green crop feature extraction | |
| CN108710840B (en) | Visual navigation path identification method for farmland pesticide spraying robot | |
| CN102214306B (en) | Leaf disease spot identification method and device | |
| JP2010220197A (en) | Device and method for detecting shadow in image | |
| CN110610506B (en) | Image processing technology-based agaricus blazei murill fruiting body growth parameter detection method | |
| CN106127735B (en) | A kind of facilities vegetable edge clear class blade face scab dividing method and device | |
| Wang et al. | Combining SUN-based visual attention model and saliency contour detection algorithm for apple image segmentation | |
| CN111798470A (en) | Crop image entity segmentation method and system applied to intelligent agriculture | |
| CN111008642A (en) | High-resolution remote sensing image classification method and system based on convolutional neural network | |
| CN115862004A (en) | Corn ear surface defect detection method and device | |
| CN115861686A (en) | Litchi key growth period identification and detection method and system based on edge deep learning | |
| US11880981B2 (en) | Method and system for leaf age estimation based on morphological features extracted from segmented leaves | |
| CN111079637A (en) | Method, device and equipment for segmenting rape flowers in field image and storage medium | |
| CN107239761B (en) | Fruit tree branch pulling effect evaluation method based on skeleton angular point detection | |
| CN111612797B (en) | Rice image information processing system | |
| Tran et al. | Automatic dragon fruit counting using adaptive thresholds for image segmentation and shape analysis | |
| CN115601690B (en) | Edible fungus environment detection method based on intelligent agriculture | |
| Nawawi et al. | Comprehensive pineapple segmentation techniques with intelligent convolutional neural network | |
| CN111401121A (en) | Method for realizing citrus segmentation based on super-pixel feature extraction |
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 |