CN115271200B - Intelligent coherent picking system for famous tea - Google Patents
Intelligent coherent picking system for famous tea Download PDFInfo
- Publication number
- CN115271200B CN115271200B CN202210876003.6A CN202210876003A CN115271200B CN 115271200 B CN115271200 B CN 115271200B CN 202210876003 A CN202210876003 A CN 202210876003A CN 115271200 B CN115271200 B CN 115271200B
- Authority
- CN
- China
- Prior art keywords
- module
- picking
- tea
- coordinates
- buds
- 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
- 241001122767 Theaceae Species 0.000 title claims abstract description 118
- 230000001427 coherent effect Effects 0.000 title claims abstract description 8
- 230000003993 interaction Effects 0.000 claims abstract description 6
- 238000004891 communication Methods 0.000 claims abstract description 4
- 238000000034 method Methods 0.000 claims description 23
- 239000011159 matrix material Substances 0.000 claims description 22
- 239000012636 effector Substances 0.000 claims description 17
- 238000010008 shearing Methods 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 14
- 238000001514 detection method Methods 0.000 claims description 9
- 230000004927 fusion Effects 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 8
- 230000007246 mechanism Effects 0.000 claims description 7
- 238000012216 screening Methods 0.000 claims description 7
- 230000006870 function Effects 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 230000002776 aggregation Effects 0.000 claims description 5
- 238000004220 aggregation Methods 0.000 claims description 5
- 230000002457 bidirectional effect Effects 0.000 claims description 5
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000012847 principal component analysis method Methods 0.000 claims description 4
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 230000007786 learning performance Effects 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 230000000007 visual effect Effects 0.000 abstract description 6
- 238000000513 principal component analysis Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 238000005286 illumination Methods 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 241000540325 Prays epsilon Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000002262 irrigation Effects 0.000 description 1
- 238000003973 irrigation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Life Sciences & Earth Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Agronomy & Crop Science (AREA)
- Animal Husbandry (AREA)
- Mining & Mineral Resources (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Marine Sciences & Fisheries (AREA)
- Operations Research (AREA)
- Primary Health Care (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Processing (AREA)
Abstract
The invention provides an intelligent coherent picking system for famous tea, which comprises a target identification module, a path planning module, a picking execution module and an operation control module; the operation control module is respectively in communication interaction with the target identification module, the path planning module and the picking execution module. The target identification module obtains picking points by constructing a tea tender bud image data set, improving a YoloV5 network model and fitting a minimum external cuboid of a three-dimensional point cloud of tea tender buds. The system realizes accurate positioning of tea buds through the visual recognition system, and simultaneously realizes intelligent and mechanized tea picking through the picking robot without a large amount of labor force, and has high picking efficiency and high famous tea yield.
Description
Technical Field
The invention relates to the technical field of tea picking, in particular to an intelligent coherent picking system for famous tea.
Background
In recent years, the planting and production of famous and excellent tea are greatly developed in China, so that the income of tea farmers is obviously increased. The famous tea is picked by a manual or handheld tea picking machine, so that the manual picking selectivity is high, the picking accuracy is high, the damage to the tea is not easy to occur, but the manual picking wastes a large amount of labor force, and the picking efficiency is low; the handheld picking machine is low in cost, high in efficiency and capable of effectively saving labor force, but most of handheld picking machines adopt a one-knife reciprocating cutting mode, and tea leaves are influenced by illumination, temperature and humidity, irrigation modes and the like, the growth heights of tea leaves buds are different, meanwhile, the volumes of the tea leaves buds are small, the forms are different, the space distribution is dense, and the one-knife reciprocating cutting mode is extremely easy to cause the problems of missing picking, wrong picking, even breakage of the tea leaves and the like. In addition, because the tea layers obtained by the existing handheld tea picking machine are uneven, subsequent classification and screening still need to be manually carried out, and the labor cost and the time of the tea picking process are increased.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention aims to provide an intelligent coherent picking system for famous tea, which realizes accurate positioning of tea buds through a visual identification system, and simultaneously realizes intelligent and mechanical tea picking through a picking robot, so that a large amount of labor force is not needed, the picking efficiency is high, and the famous tea yield is high.
The aim of the invention is achieved by the following technical scheme:
an intelligent coherent picking system for famous tea, which is characterized in that: the picking device comprises a target identification module, a path planning module, a picking execution module and an operation control module; the operation control module is respectively in communication interaction with the target identification module, the path planning module and the picking execution module;
the target identification module is used for acquiring picking points of tea buds;
the path planning module is used for planning the travelling path of the picking execution module, and specifically comprises the following steps: the operation control module inputs the picking points obtained by the target identification module and the positions of the current picking execution module into the path planning module, and the path planning module fits the three-dimensional coordinates of the picking execution module with the three-dimensional coordinates of the picking points to obtain the shortest and optimal picking path;
the picking execution module comprises a moving mechanical arm and an end effector, and the moving mechanical arm moves the end effector to each picking point according to the picking path planned in the path planning module; the end effector comprises a shearing part and a telescopic part, picking of tea buds is achieved through closing of the shearing part, and collection of tea is achieved through shrinkage of the telescopic part and opening of the shearing part.
Further optimizing, the specific steps of the target identification module for acquiring picking points of tea buds are as follows:
s10, constructing a tea bud image data set; s20, constructing a feature map with rich semantic information through a Bi-directional feature pyramid network (Bi-Directional Feature Pyramid Network, biFPN) and a channel attention mechanism (Efficient Channel Attention, ECA) in the data set of the step S10, so as to improve a YoloV5 network model, obtain an improved YoloV5 network model and realize detection of small-size tea buds; s30, obtaining a three-dimensional point cloud of the tea based on the training result of the improved YoloV5 network model in the step S20; screening out three-dimensional point clouds of tea buds from the three-dimensional point clouds of tea leaves; finally, fitting a minimum external cuboid of the tea buds to obtain the accurate positions and picking points of the tea buds;
the YoloV5 network model comprises a Backbone module, a Neck module and a Head module; the backstone module comprises a Focus module, an SPP module and a CBS module which are used for slicing the picture, and a CSP module which is used for enhancing the learning performance of the whole convolutional neural network; the Neck module comprises a CBS module and a CSP module; the Head module includes a Detect layer that utilizes mesh-based anchors for object detection on feature maps of different scales.
Preferably, the yolv 5 network model adopts a network model with the smallest model file size and the smallest depth and width of the feature map.
Further preferably, the step S10 specifically includes: firstly, acquiring image data of tea buds by using an RGB-D camera to obtain color images and depth images of the tea buds; marking the color image by using a marking tool, performing data set enhancement operation, and expanding the number of the data sets to construct a tea sprout image data set; finally, dividing the data set to form a training set, a testing set and a verification set;
the step S20 specifically includes:
s21, preprocessing the images in the training set in the step S10, and unifying the resolutions of all the images in the training set; inputting the preprocessed image into a Backbone module to obtain feature images with different sizes; s22, inputting the feature graphs with different sizes in the step S21 into a Neck module, and adopting a bidirectional feature pyramid network to replace an original path aggregation network (Path Aggregation Network, PANet) in the Neck module to perform multi-feature fusion; then up-sampling and down-sampling the feature images in sequence, generating feature images with various sizes through the splicing of a channel attention mechanism, and inputting the feature images into a Detect layer of a Head module; s23, carrying out back propagation by combining a plurality of loss functions, and updating and adjusting weight parameters for gradients in the model; and S24, verifying the existing model by adopting the verification set in the step S10 to obtain the improved YoloV5 network model.
Preferably, the labeling tool is a Labelimg labeling tool.
Further preferably, the step S30 specifically includes:
s31, acquiring detection frame coordinates according to the result of the improved YoloV5 network model in the step S20, and generating a color image and a region of interest (Region of Interest, ROI) of a corresponding depth image;
s32, according to the mapping relation between the pixel coordinates of the depth image and the pixel coordinates of the color image, obtaining corresponding mapped color image coordinates through the coordinate values, the pixel values and the recording distances of the depth image;
s33, obtaining a three-dimensional point cloud of the tea through coordinate fusion of the color image and the depth image, wherein the three-dimensional point cloud comprises the following specific steps of:
in the method, in the process of the invention,a coordinate system representing a three-dimensional point cloud; />A coordinate system representing a color image; d represents a depth value, obtained by a depth image; f (f) x 、f y Representing the camera focal length;
s34, because the generated three-dimensional point cloud of the tea leaves comprises tea leaves buds and background point clouds thereof, an average value of the three-dimensional point cloud of the tea leaves is obtained through calculation, and the average value is used as a distance threshold value; filtering the background point cloud which is larger than the distance threshold value to obtain a primarily segmented three-dimensional point cloud; adopting a DBSCAN clustering algorithm, and setting a parameter radius Eps and the minimum sample number M required to be contained in the neighborhood p The primarily segmented three-dimensional point clouds are gathered into one type, and three-dimensional point clouds of tea buds are screened out;
s35, adopting a principal component analysis (Principal Component Analysis, PCA) to fit the minimum external cuboid of the tea buds at the position according to the growth posture of the tea buds; calculating to obtain the coordinates of each vertex of the cuboid; and obtaining the coordinates of the center point of the bottom surface of the cuboid by solving the average value of four vertexes of the bottom surface of the cuboid, and taking the point as a picking point of tea buds.
Further preferably, the step S35 specifically includes:
s351, screening three main directions of the three-dimensional point cloud of the tea buds, namely x, y and z directions by adopting a principal component analysis method, and calculating mass centers and covariance to obtain a covariance matrix; the method comprises the following steps:
wherein P is c Representing centroid coordinates of the three-dimensional point cloud; n represents the number of three-dimensional point clouds (i.e., the number of points); (x) i ,y i ,z i ) Representing three-dimensional coordinates of the i-th point;
wherein C is p Representing a covariance matrix of the three-dimensional point cloud;
s352, singular value decomposition is carried out on the covariance matrix to obtain a characteristic value and a characteristic vector, wherein the specific formula is as follows:
in U p Representing covariance matrix C p C p T Is a feature vector matrix of (a); d (D) p Indicating that a non-0 value on a diagonal is C p C p T A diagonal matrix of square roots of non-0 eigenvalues;represents a C p T C p Is a feature vector matrix of (a);
the direction of the feature vector corresponding to the maximum feature value is the main axis direction of the cuboid;
s353, projecting coordinate points onto the direction vector, calculating the position coordinates P of each vertex i Interior to coordinate point unit vectorObtaining the maximum value and the minimum value of the product in each direction, enabling a, b and c to be the average value of the maximum value and the minimum value of x, y and z respectively, obtaining the center point O and the length L of the cuboid, and generating the cuboid with the most proper and compact tea bud;
the specific formula is as follows:
O=ax+by+cz;
wherein X is a unit vector of the coordinate point in the X direction; y is a unit vector of the coordinate point in the Y direction; z is a unit vector of the coordinate point in the Z direction; l (L) x 、L y 、L z The lengths of the cuboids in the x direction, the y direction and the z direction are respectively;
s354, judging the coordinates of the smallest four points in the y direction of the cuboid to be used as four vertex coordinates of the bottom surface of the cuboid; and finally, obtaining the coordinate of the center point of the bottom surface of the cuboid, namely the picking point, through the average value of the four vertex coordinates.
The invention has the following technical effects:
according to the method, the small targets of the tea shoots under the large visual field are checked and positioned through the deep learning method, firstly, the detection model of the small-size tea shoots is constructed, and meanwhile, the semantic expression and positioning capability of the image on multiple scales are enhanced through the improved YoloV5 network model, so that the method is suitable for judging and identifying the tea shoots with small targets, different forms and dense distribution space, misjudgment and missed judgment caused by the difference or mutual overlapping of the tea shoots in the tea shoot identification process are avoided, and the accuracy of identification and judgment is improved. Simultaneously, this application is through the minimum external cuboid of fitting tealeaves tender bud, realize the accurate positioning of tealeaves picking the point (use minimum external cuboid bottom surface central point to be the picking point of tealeaves tender bud promptly), effectively avoid picking the damage that causes tealeaves tender bud, ensure simultaneously that the whole tealeaves tender bud of most effective, complete picking, ensure the quality of tealeaves. Through the cooperation of the target identification module, the path planning module, the picking execution module and the operation control module, the shortest and optimal picking feeding path can be obtained efficiently, the picking efficiency is improved, and the picking flow is simplified, so that automatic and mechanical tea picking is realized, the use of labor force is reduced, and the time and cost for tea picking are saved.
Drawings
Fig. 1 is a flow chart of a picking system for picking tea leaves in an embodiment of the invention.
Fig. 2 is a flowchart of a target recognition module according to an embodiment of the present application obtaining picking points.
Fig. 3 is a schematic diagram of a picture marked by using a marking tool in the embodiment of the present application.
Fig. 4 is a multi-scale feature fusion structure diagram based on a bi-directional feature pyramid network structure in an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Examples:
an intelligent coherent picking system for famous tea, which is characterized in that: the picking device comprises a target identification module, a path planning module, a picking execution module and an operation control module; the operation control module is respectively communicated and interacted with the target identification module, the path planning module and the picking execution module;
the target identification module is used for acquiring picking points of tea buds, and specifically comprises the following steps:
s10, constructing a tea bud image data set: firstly, acquiring image data of tea buds by using an RGB-D camera to obtain color images and depth images of the tea buds; then, marking the color image by using a marking tool, such as a Labelimg marking tool (shown in figure 1), performing dataset enhancement operation (the dataset enhancement operation can be realized by adopting the conventional technical means, such as space conversion, color conversion and the like, and expanding the number of datasets to construct a tea shoot image dataset; and finally, dividing the data set to form a training set, a testing set and a verification set.
S20, constructing a feature map with rich semantic information through a Bi-directional feature pyramid network (Bi-Directional Feature Pyramid Network, biFPN) and a channel attention mechanism (Efficient Channel Attention, ECA) in the data set of the step S10, so as to improve a YoloV5 network model, obtain an improved YoloV5 network model and realize detection of small-size tea buds;
the YoloV5 network model adopts a network model with the minimum model file size and the minimum depth and width of a feature map, and comprises a Backbone module, a Neck module and a Head module; the backstone module comprises a Focus module, an SPP module and a CBS module which are used for slicing the picture, and a CSP module which is used for enhancing the learning performance of the whole convolutional neural network; the Neck module comprises a CBS module and a CSP module; the Head module comprises a Detect layer for detecting targets on feature graphs with different scales by using a grid-based anchor;
the method specifically comprises the following steps:
s21, preprocessing the images in the training set in the step S10, and unifying the resolutions of all the images in the training set; inputting the preprocessed image into a Backbone module to obtain feature images with different sizes;
s22, inputting the feature graphs with different sizes in the step S21 into a Neck module, and adopting a bidirectional feature pyramid network to replace an original path aggregation network (Path Aggregation Network, PANet) in the Neck module to perform multi-feature fusion; then up-sampling and down-sampling the feature images in sequence, generating feature images with various sizes through the splicing of a channel attention mechanism, and inputting the feature images into a Detect layer of a Head module;
in the yolv5 network model (i.e., in the existing yolv5 network structure), the reinforcing feature is used to extract BiFPN, upsample p5_in, and then, after upsampling, bifpn_concat stacking with p4_in to obtain p4_td; then upsampling is carried out on the P4_td, and BiFPN_Concat stacking is carried out on the upsampled P3_in to obtain P3_out; then, carrying out downsampling on the P3_out, and carrying out BiFPN_Concat stacking with the P4_td after downsampling to obtain the P4_out; and then downsampling is carried out on the P4_out, and the downsampled P5_out is stacked with the P5_in to obtain the P5_out. According to the method, efficient bidirectional cross connection is used for feature fusion, nodes with small contribution to feature fusion in PANet are removed, additional connection is added between input nodes and output nodes at the same level, more features are fused without additional cost, and semantic expression and positioning capability on multiple scales are enhanced, as shown in figure 2.
Then adding ECA after the 9 th layer, the module changes the input feature map from a matrix of [ h, w, c ] to a vector of [1, c ] through global average pooling (Global Average Pooling), then calculates to obtain an adaptive one-dimensional convolution kernel size, and uses the kernel size in one-dimensional convolution to obtain the weight of each channel in the feature map; and multiplying the normalized weight and the original input feature map channel by channel to generate a weighted feature map.
The attention mechanism uses a 1x1 convolution layer after the global average pooling layer, so that a full connection layer is removed, dimension reduction is avoided, cross-channel interaction is effectively captured, and finally the probability of judging an object and the detection precision of a model are improved; the specific formula is as follows:
wherein, C represents the channel dimension; k represents volume and size; y and b are respectively 2 and 1;
s23, carrying out back propagation by combining various loss functions (such as classification loss, positioning loss, execution loss and the like), and updating and adjusting weight parameters for gradients in the model;
and S24, verifying the existing model by adopting the verification set in the step S10 to obtain the improved YoloV5 network model.
S30, obtaining a three-dimensional point cloud of the tea based on the training result of the improved YoloV5 network model in the step S20; screening out three-dimensional point clouds of tea buds from the three-dimensional point clouds of tea leaves; finally, fitting a minimum external cuboid of the tea buds to obtain the accurate positions and picking points of the tea buds; the method comprises the following steps:
s31, firstly, according to the result of the improved YoloV5 network model in the step S20, acquiring detection frame coordinates, and generating a color image and a region of interest (Region of Interest, ROI) of a corresponding depth image;
s32, obtaining corresponding mapped color image coordinates according to the mapping relation between the pixel coordinates of the depth image and the pixel coordinates of the color image and through the coordinate values, the pixel values and the recording distances of the depth image;
s33, obtaining a three-dimensional point cloud of the tea through coordinate fusion of the color image and the depth image, wherein the three-dimensional point cloud comprises the following specific steps of:
in the method, in the process of the invention,a coordinate system representing a three-dimensional point cloud; />A coordinate system representing a color image; d represents a depth value, obtained by a depth image; f (f) x 、f y Representing the camera focal length;
s34, because the generated three-dimensional point cloud of the tea leaves comprises tea leaves buds and background point clouds thereof, an average value of the three-dimensional point cloud of the tea leaves is obtained through calculation, and the average value is used as a distance threshold value; filtering the background point cloud which is larger than the distance threshold value to obtain a primarily segmented three-dimensional point cloud; adopting a DBSCAN clustering algorithm, and setting a parameter radius Eps and the minimum sample number M required to be contained in the neighborhood p The primarily segmented three-dimensional point clouds are gathered into one type, and three-dimensional point clouds of tea buds are screened out;
wherein, the DBSCAN clustering algorithm randomly selects a data sample in space, and determines whether the sample number distributed in the neighborhood radius Eps is greater than or equal to the minimum sample number M p The threshold number determines whether it is a core object:
if so, dividing all points in the neighborhood into the same cluster group, and meanwhile, on the basis of the cluster group above, searching for all samples with reachable density by breadth-first search, and dividing the samples into the cluster group;
if the data sample is a non-core object, marking the data sample as noise point removal;
the formula is specifically as follows:
N Eps (p)={q∈D|dist(p,q)≤Eps};
wherein D represents a point cloud sample set; p and q respectively represent sample points summarized by the sample set;
for any p.epsilon.D, if its Eps corresponds to |N Eps (p) | contains at least M p P is the core object; if q is within Eps of p and p is the core object, then q is reachable from p density;
s35, adopting a principal component analysis (Principal Component Analysis, PCA) to fit the minimum external cuboid of the tea buds at the position according to the growth posture of the tea buds; calculating to obtain the coordinates of each vertex of the cuboid; obtaining the coordinates of the center point of the bottom surface of the cuboid by solving the average value of four vertexes of the bottom surface of the cuboid, and taking the point as a picking point of tea buds; the method comprises the following steps:
s351, screening three main directions of the three-dimensional point cloud of the tea buds, namely x, y and z directions by adopting a principal component analysis method, and calculating mass centers and covariance to obtain a covariance matrix; the method comprises the following steps:
wherein P is c Representing centroid coordinates of the three-dimensional point cloud; n represents the number of three-dimensional point clouds (i.e., the number of points); (x) i ,y i ,z i ) Representing three-dimensional coordinates of the i-th point;
wherein C is p Representing a covariance matrix of the three-dimensional point cloud;
s352, singular value decomposition is carried out on the covariance matrix to obtain a characteristic value and a characteristic vector, wherein the specific formula is as follows:
in U p Representing covariance matrix C p C p T Is a feature vector matrix of (a); d (D) p Indicating that a non-0 value on a diagonal is C p C p T A diagonal matrix of square roots of non-0 eigenvalues;represents a C p T C p Is a feature vector matrix of (a);
the direction of the feature vector corresponding to the maximum feature value is the main axis direction of the cuboid;
s353, projecting coordinate points onto the direction vector, calculating the position coordinates P of each vertex i Obtaining the maximum value and the minimum value of the inner product of the coordinate point unit vector in each direction, enabling a, b and c to be the average value of the maximum value and the minimum value on x, y and z respectively, obtaining the center point O and the length L of the cuboid, and generating the cuboid with the most proper and compact tea bud;
the specific formula is as follows:
O=ax+by+cz;
wherein X is a unit vector of the coordinate point in the X direction; y is a unit vector of the coordinate point in the Y direction; z is a unit vector of the coordinate point in the Z direction; l (L) x 、L y 、L z The lengths of the cuboids in the x direction, the y direction and the z direction are respectively;
s354, judging the coordinates of the smallest four points in the y direction of the cuboid to be used as four vertex coordinates of the bottom surface of the cuboid; and finally, obtaining the coordinate of the center point of the bottom surface of the cuboid, namely the picking point, through the average value of the four vertex coordinates.
The path planning module is used for planning the travelling path of the picking execution module, and specifically comprises the following steps: the operation control module inputs the positions of each picking point and the current picking execution module (specifically, the shearing part) obtained by the target identification module to the path planning module, and the path planning module fits the three-dimensional coordinates of the picking execution module (specifically, the shearing part) with the three-dimensional coordinates of each picking point to obtain the shortest and optimal picking path;
for example: a, a 1 Constructing picking execution module (specifically, shear part) picking sequence path objective function: firstly, modeling picking sequence path planning of a picking execution module (particularly a shear part) based on Markov decision process theory (Markov Decision Processes, MDP); then, the operation control module controls the picking execution module (in particular the shearing part) to execute a prediction action according to the received three-dimensional information of the picking points obtained by the method in the target identification module, and then observes a new state and obtains rewards, so that the target function maximizes the expected accumulated rewards obtained by the picking execution module (in particular the shearing part); a, a 2 Setting up a simulation environment containing a tea plantation and a picking execution module by using simulation equipment, and combining illumination intensity and illumination intensityPhysical quantities such as machine direction, tea bud position, color and the like are used as parameters of a simulation environment for training; in the training process, the randomness of the simulation environment is gradually increased, data with continuously increased learning difficulty is acquired through the interactive acquisition of the robot and the environment, and the data is sampled; a, a 3 And optimizing the objective function by adopting a near-end strategy optimization algorithm module (Proximal Policy Optimization, PPO) and combining a solver to obtain a picking sequence planning model of a picking execution module (particularly a shearing part).
The picking execution module comprises a moving mechanical arm and an end effector, wherein the moving mechanical arm moves the end effector to each picking point according to a picking path planned in the path planning module, and the moving mechanical arm can adopt a multi-degree-of-freedom mechanical arm to select the degree of freedom of the mechanical arm according to actual picking conditions; and communication interaction is realized between the motion mechanical arm and the end effector through the operation control module. The end effector comprises a shearing part and a telescopic part, picking of tea buds is achieved through closing of the shearing part, and collection of tea is achieved through shrinkage of the telescopic part and opening of the shearing part. If necessary, a tea collecting basket is arranged on the moving mechanical arm and positioned on the end effector.
The operation control module can build Ubuntu18.04 system on the edge computer and simultaneously create a Pytorch-based deep learning operation environment; deploying the target identification module and the path planning module into an edge computer, and cooperatively managing resource scheduling and network signal interaction of the two neural network modules; while deploying the motion robot control system and the end effector drive system in the edge computer.
The specific method for picking tea leaves of the picking system comprises the following steps:
s001, initializing a system in a tea garden, and then manually controlling a motion mechanical arm to drive an end effector to move to a pre-tea-picking point;
s002, the operation control module opens the target recognition module and the path planning module and simultaneously opens the camera;
s003, the camera transmits the visual field to the target recognition module, and the target recognition module carries out multi-target detection on the tea buds in the image; detecting tea buds and fitting out a smallest external cuboid of the tea buds to obtain a bottom surface central point of the smallest external cuboid; then, the center point of the bottom surface of the minimum external cuboid is transmitted to an operation control module;
s004, the operation control module guides the center point of the bottom surface of the minimum circumscribed cuboid and the current position coordinate of the end effector into a PPO model of the path planning module to obtain the shortest optimal path picked by the end effector; transmitting the path obtained by the calculation to an operation control module;
s005, the operation control module transmits the path information to the picking execution module, and the picking execution module controls the end effector to reach a picking point through the motion mechanical arm, so that the end effector can realize picking and collecting actions; until picking of tea leaves and buds in the current visual field of the camera is completed;
s006, moving the picking system to the next visual field range, and continuing picking of the tea buds.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (1)
1. An intelligent coherent picking system for famous tea, which is characterized in that: the picking device comprises a target identification module, a path planning module, a picking execution module and an operation control module; the operation control module is respectively in communication interaction with the target identification module, the path planning module and the picking execution module;
the target identification module is used for acquiring picking points of tea buds;
the path planning module is used for planning the travelling path of the picking execution module, and specifically comprises the following steps: the operation control module inputs the picking points obtained by the target identification module and the positions of the current picking execution module into the path planning module, and the path planning module fits the three-dimensional coordinates of the picking execution module with the three-dimensional coordinates of the picking points to obtain the shortest and optimal picking path;
the picking execution module comprises a moving mechanical arm and an end effector, and the moving mechanical arm moves the end effector to each picking point according to the picking path planned in the path planning module; the end effector comprises a shearing part and a telescopic part, picking of tea buds is realized through closing of the shearing part, and collection of tea is realized through shrinkage of the telescopic part and opening of the shearing part;
the specific steps of the target identification module for acquiring picking points of tea buds are as follows:
s10, constructing a tea bud image data set; s20, constructing a feature map with rich semantic information through a bidirectional feature pyramid network and a channel attention mechanism in the data set of the step S10, improving a YoloV5 network model, obtaining an improved YoloV5 network model and detecting small-size tea buds; s30, obtaining a three-dimensional point cloud of the tea based on the training result of the improved YoloV5 network model in the step S20; screening out three-dimensional point clouds of tea buds from the three-dimensional point clouds of tea leaves; finally, fitting a minimum external cuboid of the tea buds to obtain the accurate positions and picking points of the tea buds;
the YoloV5 network model comprises a Backbone module, a Neck module and a Head module; the backstone module comprises a Focus module, an SPP module and a CBS module which are used for slicing the picture, and a CSP module which is used for enhancing the learning performance of the whole convolutional neural network; the Neck module comprises a CBS module and a CSP module; the Head module comprises a Detect layer for detecting targets on feature graphs with different scales by using a grid-based anchor;
the step S10 specifically includes: firstly, acquiring image data of tea buds by using an RGB-D camera to obtain color images and depth images of the tea buds; marking the color image by using a marking tool, performing data set enhancement operation, and expanding the number of the data sets to construct a tea sprout image data set; finally, dividing the data set to form a training set, a testing set and a verification set;
the step S20 specifically includes:
s21, preprocessing the images in the training set in the step S10, and unifying the resolutions of all the images in the training set; inputting the preprocessed image into a Backbone module to obtain feature images with different sizes; s22, inputting the feature graphs with different sizes in the step S21 into a Neck module, and adopting a bidirectional feature pyramid network to replace an original path aggregation network in the Neck module to perform multi-feature fusion; then up-sampling and down-sampling the feature images in sequence, generating feature images with various sizes through the splicing of a channel attention mechanism, and inputting the feature images into a Detect layer of a Head module; s23, carrying out back propagation by combining a plurality of loss functions, and updating and adjusting weight parameters for gradients in the model; s24, verifying the existing model by adopting the verification set in the step S10 to obtain an improved YoloV5 network model;
the step S30 specifically includes:
s31, acquiring detection frame coordinates according to the result of the improved YoloV5 network model in the step S20, and generating a color image and an interested region of a corresponding depth image;
s32, according to the mapping relation between the pixel coordinates of the depth image and the pixel coordinates of the color image, obtaining corresponding mapped color image coordinates through the coordinate values, the pixel values and the recording distances of the depth image;
s33, obtaining a three-dimensional point cloud of the tea through coordinate fusion of the color image and the depth image, wherein the three-dimensional point cloud comprises the following specific steps of:
in the method, in the process of the invention,a coordinate system representing a three-dimensional point cloud; />A coordinate system representing a color image; d represents a depth value, obtained by a depth image; f (f) x 、f y Representing the camera focal length;
s34, because the generated three-dimensional point cloud of the tea leaves comprises tea leaves buds and background point clouds thereof, an average value of the three-dimensional point cloud of the tea leaves is obtained through calculation, and the average value is used as a distance threshold value; filtering the background point cloud which is larger than the distance threshold value to obtain a primarily segmented three-dimensional point cloud; adopting a DBSCAN clustering algorithm, and setting a parameter radius Eps and the minimum sample number M required to be contained in the neighborhood p The primarily segmented three-dimensional point clouds are gathered into one type, and three-dimensional point clouds of tea buds are screened out;
s35, fitting the minimum external cuboid of the tea buds at the position according to the growth posture of the tea buds by adopting a principal component analysis method; calculating to obtain the coordinates of each vertex of the cuboid; obtaining the coordinates of the center point of the bottom surface of the cuboid by solving the average value of four vertexes of the bottom surface of the cuboid, and taking the point as a picking point of tea buds;
the step S35 specifically includes:
s351, screening three main directions of the three-dimensional point cloud of the tea buds, namely x, y and z directions by adopting a principal component analysis method, and calculating mass centers and covariance to obtain a covariance matrix; the method comprises the following steps:
wherein P is c Representing centroid coordinates of the three-dimensional point cloud; n represents the number of three-dimensional point clouds (i.e., the number of points); (x) i ,y i ,z i ) Representing three-dimensional coordinates of the i-th point;
wherein C is p Representing a three-dimensional point cloudA covariance matrix;
s352, singular value decomposition is carried out on the covariance matrix to obtain a characteristic value and a characteristic vector, wherein the specific formula is as follows:
in U p Representing covariance matrix C p C p T Is a feature vector matrix of (a); d (D) p Indicating that a non-0 value on a diagonal is C p C p T A diagonal matrix of square roots of non-0 eigenvalues;represents a C p T C p Is a feature vector matrix of (a);
the direction of the feature vector corresponding to the maximum feature value is the main axis direction of the cuboid;
s353, projecting coordinate points onto the direction vector, calculating the position coordinates P of each vertex i Obtaining the maximum value and the minimum value of the inner product of the coordinate point unit vector in each direction, enabling a, b and c to be the average value of the maximum value and the minimum value on x, y and z respectively, obtaining the center point O and the length L of the cuboid, and generating the cuboid with the most proper and compact tea bud;
the specific formula is as follows:
O=ax+by+cz;
wherein X is a unit vector of the coordinate point in the X direction; y is a unit vector of the coordinate point in the Y direction; z is a unit vector of the coordinate point in the Z direction; l (L) x 、L y 、L z The lengths of the cuboids in the x direction, the y direction and the z direction are respectively;
s354, judging the coordinates of the smallest four points in the y direction of the cuboid to be used as four vertex coordinates of the bottom surface of the cuboid; and finally, obtaining the coordinate of the center point of the bottom surface of the cuboid, namely the picking point, through the average value of the four vertex coordinates.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210876003.6A CN115271200B (en) | 2022-07-25 | 2022-07-25 | Intelligent coherent picking system for famous tea |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210876003.6A CN115271200B (en) | 2022-07-25 | 2022-07-25 | Intelligent coherent picking system for famous tea |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115271200A CN115271200A (en) | 2022-11-01 |
CN115271200B true CN115271200B (en) | 2023-05-30 |
Family
ID=83770408
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210876003.6A Active CN115271200B (en) | 2022-07-25 | 2022-07-25 | Intelligent coherent picking system for famous tea |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115271200B (en) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113807215A (en) * | 2021-08-31 | 2021-12-17 | 贵州大学 | Tea tender shoot grading method combining improved attention mechanism and knowledge distillation |
CN216626695U (en) * | 2021-12-17 | 2022-05-31 | 昭平县农业农村局 | Wild tea leaf collection system |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108182374A (en) * | 2016-12-08 | 2018-06-19 | 广州映博智能科技有限公司 | A kind of picking point recognition methods for fruit string |
WO2018157286A1 (en) * | 2017-02-28 | 2018-09-07 | 深圳市大疆创新科技有限公司 | Recognition method and device, and movable platform |
CN111784764A (en) * | 2020-05-28 | 2020-10-16 | 西南石油大学 | Tea tender shoot identification and positioning algorithm |
CN111914715B (en) * | 2020-07-24 | 2021-07-16 | 廊坊和易生活网络科技股份有限公司 | Intelligent vehicle target real-time detection and positioning method based on bionic vision |
CN113674226A (en) * | 2021-07-31 | 2021-11-19 | 河海大学 | Tea leaf picking machine tea leaf bud tip detection method based on deep learning |
CN114332697A (en) * | 2021-12-19 | 2022-04-12 | 西安科技大学 | Method, system, equipment and medium for detecting faults of multiple types of targets in power transmission line |
CN114731840B (en) * | 2022-04-07 | 2022-12-27 | 仲恺农业工程学院 | Double-mechanical-arm tea picking robot based on machine vision |
-
2022
- 2022-07-25 CN CN202210876003.6A patent/CN115271200B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113807215A (en) * | 2021-08-31 | 2021-12-17 | 贵州大学 | Tea tender shoot grading method combining improved attention mechanism and knowledge distillation |
CN216626695U (en) * | 2021-12-17 | 2022-05-31 | 昭平县农业农村局 | Wild tea leaf collection system |
Also Published As
Publication number | Publication date |
---|---|
CN115271200A (en) | 2022-11-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3639241B1 (en) | Voxel based ground plane estimation and object segmentation | |
EP3405845B1 (en) | Object-focused active three-dimensional reconstruction | |
CN113920474B (en) | Internet of things system and method for intelligently supervising citrus planting situation | |
CN113160062B (en) | Infrared image target detection method, device, equipment and storage medium | |
CN115272791B (en) | YoloV 5-based multi-target detection and positioning method for tea leaves | |
CN115187803B (en) | Positioning method for picking process of famous tea tender shoots | |
Magistri et al. | Contrastive 3D shape completion and reconstruction for agricultural robots using RGB-D frames | |
CN111831010A (en) | Unmanned aerial vehicle obstacle avoidance flight method based on digital space slice | |
Li et al. | Development and field evaluation of a robotic harvesting system for plucking high-quality tea | |
CN115376125A (en) | Target detection method based on multi-modal data fusion and in-vivo fruit picking method based on target detection model | |
CN114120067A (en) | Object identification method, device, equipment and medium | |
CN114689038A (en) | Fruit detection positioning and orchard map construction method based on machine vision | |
CN115271200B (en) | Intelligent coherent picking system for famous tea | |
CN116138036B (en) | Secondary positioning method for picking young buds of famous tea | |
CN116704497B (en) | Rape phenotype parameter extraction method and system based on three-dimensional point cloud | |
Mejia et al. | Strawberry localization in a ridge planting with an autonomous rover | |
CN116739739A (en) | Loan amount evaluation method and device, electronic equipment and storage medium | |
CN113723833B (en) | Method, system, terminal equipment and storage medium for evaluating quality of forestation actual results | |
CN110909887B (en) | Model optimization method and device | |
CN116935235B (en) | Fresh tea leaf identification method and related device based on unmanned tea picking machine | |
CN113932712B (en) | Melon and fruit vegetable size measurement method based on depth camera and key points | |
Kiktev et al. | Flower Detection and Counting Using CNN for Thinning Decisions in Apple Trees. | |
TWI709111B (en) | Method for rapidly positioning crops | |
Velasquez | 3D segmentation and localization using visual cues in uncontrolled environments | |
Luo et al. | 3D Semantic Segmentation for Grape Bunch Point Cloud Based on Feature Enhancement |
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 |