Algorithm module configuration method, algorithm execution component and robot system
By adopting the algorithm module configuration method based on directed acyclic graph in intelligent grab technology, the construction and maintenance of complex logic in intelligent grab technology is simplified, and the logic construction framework in the existing technology is difficult to support multiple scenarios and multiple categories of grabs. The problem of getting the code improves the readability and maintainability of the code.
Patent Information
- Application Number
- CN202410956848.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2024-06-21
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-08
AI Technical Summary
When existing intelligent crawling technology deals with complex multi-scenario and multi-category crawling, the logical construction framework is difficult to support, resulting in difficulty in reusing code, modifying and problem-solving, and weak visual comprehensibility and high learning cost.
Using the algorithm module configuration method based on directed acyclic graph, the algorithm module is generated by atomizing nodes and operator units, and the input and output nodes are configured using mapping parameters to realize the mapping of data flow direction and processing logic, simplifying the input of the algorithm module. and output port configuration, and optimize the module structure by combining input and output ports.
It improves the readability and simplicity of the algorithm module, reduces the complexity of the code, simplifies code modification and problem investigation, reduces dependence on manual intervention, and improves the comprehensibility and maintainability of the logic.
Smart Images

Figure CN118916009A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to an algorithm module configuration method, an algorithm execution component and a vision-based robot system. Background Art
[0002] In order to reduce heavy manual labor and reduce possible injury risks, intelligent operation technologies (such as intelligent grasping technology) have gradually been widely used in the field of industrial manufacturing.
[0003] Intelligent grasping technology uses intelligent methods to realize the automation of grasping and various grasping operations, such as palletizing, depalletizing, etc. To realize the automation of grasping, a lot of code needs to be written.
[0004] In order to cope with the increasingly heavy and redundant code construction and the difficulty in simply reusing code, low-code has become a more common solution. Through a simple graphical interface and using drag-and-drop connections, the logical construction of code can be easily achieved.
[0005] However, when dealing with more complex logic, the results generated by such solutions are often complex and difficult for ordinary engineers to understand. At the same time, due to the complex logic, the overall low-code display will still be relatively large, making it difficult to modify and troubleshoot problems.
[0006] In the field of intelligent grasping technology, for complex multi-scenario and multi-category grasping, the existing logic construction framework is difficult to support multiple scenarios and multiple grasping objects, and requires complex processing procedures. It is still not easy to implement and needs to rely on more manual intervention and code writing. Moreover, the existing visual logic construction framework has weak visual comprehension when processing more complex logic, and the interactive logic is difficult to understand, and the learning cost is high. Summary of the invention
[0007] The present disclosure provides an algorithm module configuration method, comprising: configuring a plurality of operator units and data flows between the operator units based on a plurality of atomic nodes of a directed acyclic graph and directed edges between the atomic nodes, generating an operator unit sequence of the algorithm module, wherein the operator unit comprises an input end, an output end, and a processing logic configured between the input end and the output end;
[0008] Generate a first mapping parameter based on the processing logic name and input terminal configuration parameter of the input-side operator unit of the operator unit sequence, configure the input node of the directed acyclic graph based on the first mapping parameter, and the input node serves as the input port of the algorithm module, so that the data received by the input port can be mapped to the input terminal of the input-side operator unit based on the first mapping parameter;
[0009] A second mapping parameter is generated based on the processing logic name and output end configuration parameter of the output-side operator unit of the operator unit sequence, and the output node of the directed acyclic graph is configured based on the second mapping parameter. The output node serves as the output port of the algorithm module, so that the data output through the output end of the output-side operator unit can be mapped to the output port of the algorithm module based on the second mapping parameter.
[0010] According to the algorithm module configuration method of at least one embodiment of the present disclosure, data links are configured between the operator units of the operator unit sequence, and the data links between the operator units are configured based on the data flow between the operator units.
[0011] According to the algorithm module configuration method of at least one embodiment of the present disclosure, the atomic node is an inseparable node in the directed acyclic graph.
[0012] According to the algorithm module configuration method of at least one embodiment of the present disclosure, the input terminal configuration parameters of the operator unit include the input terminal data type and the input terminal data name, the output terminal configuration parameters of the operator unit include the output terminal data type and the output terminal data name, and the operator unit is configured with a processing logic name that characterizes the function of the processing logic of the operator unit;
[0013] The first mapping parameter is generated based on the processing logic name and input terminal configuration parameter of the input side operator unit of the operator unit sequence, including:
[0014] Extracting the processing logic name and input terminal configuration parameters of the input-side operator unit;
[0015] A mapping name of the first mapping parameter is constructed based on the input end data name and the processing logic name, the input end data type is used as the data type of the first mapping parameter, and a mapping relationship of the first mapping parameter is constructed based on the data type of the first mapping parameter and the input end data type.
[0016] According to at least one embodiment of the algorithm module configuration method of the present disclosure, generating the second mapping parameter based on the processing logic name and output terminal configuration parameter of the output side operator unit of the operator unit sequence includes:
[0017] Extracting the processing logic name and output terminal configuration parameters of the output side operator unit;
[0018] A mapping name of the second mapping parameter is constructed based on the output end data name and the processing logic name, the output end data type is used as the data type of the second mapping parameter, and a mapping relationship of the second mapping parameter is constructed based on the data type of the second mapping parameter and the output end data type.
[0019] According to the algorithm module configuration method of at least one embodiment of the present disclosure, configuring the input node of the directed acyclic graph based on the first mapping parameter includes:
[0020] The data type in the first mapping parameter is used as the input data type of the input node, the mapping name in the first mapping parameter is used as the input data name of the input node, and a mapping between the input node and the input end of the input-side operator unit is established based on the mapping relationship in the first mapping parameter.
[0021] According to the algorithm module configuration method of at least one embodiment of the present disclosure, configuring the output node of the directed acyclic graph based on the second mapping parameter includes:
[0022] The data type in the second mapping parameter is used as the output data type of the output node, the mapping name in the second mapping parameter is used as the output data name of the output node, and a mapping between the output node and the output end of the output-side operator unit is established based on the mapping relationship in the second mapping parameter.
[0023] According to the algorithm module configuration method of at least one embodiment of the present disclosure, the operator unit sequence includes a plurality of input-side operator units with the same processing logic, the same input-end data type and different input-end data names, each input-side operator unit maps an input port of the input node, and each input port is configured based on a first mapping parameter;
[0024] The configuration method further includes:
[0025] Determine multiple input-side operator units whose inputs need to be merged;
[0026] Performing uniform processing on the input-end data names of the multiple input-side operator units that need to be merged and inputted, and generating a unified input-end data name;
[0027] The multiple input ports corresponding to the multiple input-side operator units that need to merge inputs are merged into one input port, and a mapping name of the merged input port is generated based on the unified input port data name and the processing logic name of the input-side operator unit.
[0028] According to the algorithm module configuration method of at least one embodiment of the present disclosure, the operator unit sequence includes a plurality of output-side operator units with the same processing logic, the same output terminal data type and different output terminal data names, each output-side operator unit maps an output port of the output node, and each output port is configured based on a second mapping parameter;
[0029] The configuration method further includes:
[0030] Determine multiple output-side operator units that need to be combined for output;
[0031] Performing consistency processing on the output-end data names of the multiple output-side operator units that need to be combined and outputted, and generating a unified output-end data name;
[0032] The multiple output ports corresponding to the multiple output-side operator units that need to be merged for output are merged into one output port, and a mapping name of the merged output port is generated based on the unified output port data name and the processing logic name of the output-side operator unit.
[0033] According to at least one embodiment of the present disclosure, the algorithm module configuration method further includes:
[0034] An enabling switch is configured for the algorithm module based on the judgment logic, and the algorithm module can be executed based on a true value or a false value of the judgment logic received by the enabling switch.
[0035] According to another aspect of the present disclosure, an algorithm module is provided, wherein the algorithm module is an algorithm module configured based on the configuration method of any embodiment of the present disclosure.
[0036] According to yet another aspect of the present disclosure, there is provided an algorithm execution component, comprising:
[0037] A target data output module, which processes the input data and outputs the target data;
[0038] A conditional judgment module, wherein the conditional judgment module is configured with conditional judgment logic, receives the target data, and performs conditional judgment on the target data based on its own conditional judgment logic, and outputs a true value or a false value;
[0039] Two algorithm modules, one algorithm module can be executed based on the true value output by the condition judgment module, and the other algorithm module can be executed based on the false value output by the condition judgment module;
[0040] Among them, at least one of the algorithm modules is an algorithm module of any embodiment of the present disclosure.
[0041] According to yet another aspect of the present disclosure, there is provided an algorithm execution component, comprising:
[0042] A target data output module, which processes the input data and outputs the target data;
[0043] A plurality of conditional judgment modules, each of which is configured with a conditional judgment logic, the conditional judgment logics of the conditional judgment modules are different from each other, each of the conditional judgment modules receives the target data, and performs conditional judgment on the target data based on its own conditional judgment logic, and outputs a true value or a false value;
[0044] A plurality of Boolean operation units, each of which is configured with a Boolean judgment logic, the Boolean judgment logics of the Boolean operation units are different from each other, each of which is configured with a first input terminal, a second input terminal and a Boolean output terminal, the first input terminal of each Boolean operation unit receives a true value output by a condition judgment module, the second input terminal receives a true value output by another condition judgment module, each Boolean operation unit performs a Boolean judgment on the true values received by the first input terminal and the second input terminal based on its own Boolean judgment logic, and outputs a true value or a false value via the Boolean output terminal;
[0045] a plurality of algorithm modules, each algorithm module being connected to a Boolean output terminal of a Boolean operation unit and capable of being executed based on a true value or a false value output by the Boolean output terminal of the Boolean operation unit;
[0046] Wherein, at least one of the algorithm modules is the algorithm module described in claim 11.
[0047] According to the algorithm execution component of at least one embodiment of the present disclosure, two algorithm modules among the multiple algorithm modules are connected to the Boolean output end of the same Boolean operation unit, one of the two algorithm modules can be executed based on the true value output by the Boolean operation unit, and the other algorithm module can be executed based on the false value output by the Boolean operation unit.
[0048] According to the algorithm execution component of at least one embodiment of the present disclosure, the target data includes a two-dimensional image and a three-dimensional point cloud acquired by a camera.
[0049] According to the algorithm execution component of at least one embodiment of the present disclosure, the condition judgment logic is a target object category judgment logic, a distance judgment logic between the target object and the camera, or a target object direction judgment logic.
[0050] According to yet another aspect of the present disclosure, a readable storage medium is provided, in which the algorithm execution component of any one embodiment of the present disclosure is stored in the form of a computer program.
[0051] According to yet another aspect of the present disclosure, a computer program product is provided. The computer program product is an algorithm execution component of any one of the embodiments of the present disclosure.
[0052] According to another aspect of the present disclosure, there is provided a vision-based robot system, comprising:
[0053] A camera, wherein the camera is used to obtain a two-dimensional image and a three-dimensional point cloud of a target object;
[0054] The algorithm execution component of any one of the embodiments of the present disclosure processes the two-dimensional image and the three-dimensional point cloud acquired by the camera to obtain the position and posture information of the target object;
[0055] A robot operates the target object based on the position information of the target object. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] The accompanying drawings illustrate exemplary embodiments of the present disclosure and together with the description serve to explain the principles of the present disclosure. These drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification.
[0057] Figure 1 It is a flowchart of an algorithm module configuration method according to one embodiment of the present disclosure.
[0058] Figure 2 The input port and output port of the algorithm module of one embodiment of the present disclosure are exemplarily shown.
[0059] Figure 3 An operator unit sequence of one embodiment of the present disclosure is exemplarily shown.
[0060] Figure 4 The parameter configuration interface of the input port of the algorithm module (ie, subgraph) of one embodiment of the present disclosure is exemplarily shown.
[0061] Figure 5 It is a flowchart of a method for constructing a first mapping parameter according to an embodiment of the present disclosure.
[0062] Figure 6 It is a flowchart of a method for constructing a second mapping parameter according to an embodiment of the present disclosure.
[0063] Figure 7 The schematic diagram of the input nodes and output nodes of an algorithm module configured with multiple input ports and multiple output ports (before merging) is exemplarily shown.
[0064] Figure 8 Shows Figure 7 Parameter configuration interface of the algorithm module in .
[0065] Fig. 9 A schematic flow chart of an algorithm module configuration method according to another embodiment of the present disclosure is shown.
[0066] Fig.10 yes Figure 7 Schematic diagram of the algorithm module after the input nodes are merged.
[0067] Fig.11 Shows Fig.10 Parameter configuration interface of the algorithm module in .
[0068] Fig.12 A schematic structural block diagram of an algorithm execution component according to an embodiment of the present disclosure is shown.
[0069] Fig.13 An example of an algorithm execution component according to an embodiment of the present disclosure is shown.
[0070] Fig.14 A schematic structural block diagram of an algorithm execution component according to another embodiment of the present disclosure is shown.
[0071] Fig.15 It is an example of an algorithm execution component of another embodiment of the present disclosure.
[0072] Fig.16 The complete execution chain of an algorithm execution component example of the present disclosure is shown. DETAILED DESCRIPTION
[0073] The present disclosure is further described in detail below in conjunction with the accompanying drawings and implementations. It is understood that the specific implementations described herein are only used to explain the relevant content, rather than to limit the present disclosure. It should also be noted that, for ease of description, only the parts related to the present disclosure are shown in the accompanying drawings.
[0074] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present disclosure can be combined with each other. The technical solution of the present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0075] Unless otherwise specified, the exemplary embodiments / embodiments shown will be understood as providing exemplary features of various details of some ways in which the technical concept of the present disclosure can be implemented in practice. Therefore, unless otherwise specified, the features of the various embodiments / embodiments can be combined, separated, interchanged and / or rearranged without departing from the technical concept of the present disclosure.
[0076] The use of cross-hatching and / or shading in the accompanying drawings is generally used to make the boundaries between adjacent components clear. As such, unless otherwise specified, the presence or absence of cross-hatching or shading does not convey or indicate any preference or requirement for the specific materials, material properties, dimensions, proportions, commonalities between the components shown, and / or any other characteristics, attributes, properties, etc. of the components. In addition, in the accompanying drawings, the sizes and relative sizes of the components may be exaggerated for clarity and / or descriptive purposes. When the exemplary embodiments can be implemented differently, the specific process sequence can be performed in a different order than described. For example, two successively described processes can be performed substantially simultaneously or in an order opposite to the described order. In addition, the same figure numbers represent the same components.
[0077] When a component is referred to as being "on" or "over," "connected to," or "coupled to" another component, the component may be directly on, directly connected to, or directly coupled to the other component, or intervening components may be present. However, when a component is referred to as being "directly on," "directly connected to," or "directly coupled to" another component, there are no intervening components. For this purpose, the term "connected" may refer to a physical connection, an electrical connection, etc., with or without intervening components.
[0078] The terms used herein are for the purpose of describing specific embodiments, and are not intended to be restrictive. As used herein, unless the context clearly indicates otherwise, the singular forms "one (kind, person)" and "said (the)" are also intended to include plural forms. In addition, when the terms "comprise" and / or "include" and their variations are used in this specification, it is explained that there are stated features, integral bodies, steps, operations, parts, assemblies and / or their groups, but it is not excluded that there are or add one or more other features, integral bodies, steps, operations, parts, assemblies and / or their groups. It should also be noted that, as used herein, the terms "substantially", "approximately" and other similar terms are used as approximate terms and not as degree terms, so that they are used to explain the inherent deviations of the measured values, calculated values and / or the values provided that will be recognized by those of ordinary skill in the art.
[0079] Figure 1 It is a flowchart of an algorithm module configuration method according to one embodiment of the present disclosure.
[0080] refer to Figure 1 In some embodiments of the present disclosure, the algorithm module configuration method M100 of the present disclosure includes:
[0081] S102, configuring multiple operator units and data flows between the operator units based on multiple atomic nodes of a directed acyclic graph (DAG) and directed edges between the atomic nodes, generating an operator unit sequence of an algorithm module, where the operator unit includes an input end, an output end, and a processing logic configured between the input end and the output end;
[0082] S104, generating a first mapping parameter based on the processing logic name of the input-side operator unit of the operator unit sequence and the input-end configuration parameter, configuring an input node of a directed acyclic graph (DAG) based on the first mapping parameter, and the input node serves as an input port of the algorithm module, so that data received by the input port can be mapped to the input end of the input-side operator unit based on the first mapping parameter;
[0083] A second mapping parameter is generated based on the processing logic name (operator name) of the output-side operator unit of the operator unit sequence and the output-end configuration parameters, and an output node of a directed acyclic graph (DAG) is configured based on the second mapping parameter. The output node serves as an output port of the algorithm module, so that data output through the output end of the output-side operator unit can be mapped to the output port of the algorithm module based on the second mapping parameter.
[0084] The algorithm module described in the present disclosure is the algorithm subgraph in the algorithm graph (i.e., the algorithm execution component), and the processing logic configured between the input end and the output end described in the present disclosure is the processing logic (i.e., the calculation logic) implemented based on a computer program.
[0085] The input-side operator unit described in the present disclosure is the first operator unit of the operator unit sequence, and the output-side operator unit described in the present disclosure is the last operator unit of the operator unit sequence.
[0086] The processing logic name of the input-side operator unit described in the present disclosure is the operator name, and the input-end configuration parameters include the data type (also known as the parameter type) and the data name (also known as the parameter name).
[0087] The processing logic name of the output-side operator unit described in the present disclosure is the operator name, and the output-end configuration parameters include the data type (also known as the parameter type) and the data name (also known as the parameter name).
[0088] The present disclosure constructs an algorithm module based on a directed acyclic graph (DAG). First, based on multiple atomic nodes (atomic nodes, i.e., indivisible nodes) of a directed acyclic graph and directed edges between atomic nodes, multiple operator units and data flows between operator units are configured to generate an operator unit sequence of the algorithm module, which forms a complete processing logic of the algorithm module. The operator unit sequence includes an input-side operator unit and an output-side operator unit, or also includes at least one operator unit between the two. Next, a first mapping parameter is generated based on the processing logic name (i.e., operator name) of the input-side operator unit and the input-end configuration parameters, and a second mapping parameter is generated based on the processing logic name (i.e., operator name) of the output-side operator unit and the output-end configuration parameters. Finally, the input nodes of the directed acyclic graph (DAG) are configured based on the first mapping parameters, and the output nodes of the directed acyclic graph are configured based on the second mapping parameters, completing the configuration of the input port and output port of the algorithm module, so that the data received by the input port of the algorithm module can be mapped to the input end of the input side operator unit based on the first mapping parameters, and the data output through the output end of the output side operator unit can be mapped to the output port of the algorithm module based on the second mapping parameters.
[0089] Figure 2 The input port and output port of an algorithm module of an embodiment of the present disclosure are exemplarily shown, and the name of the algorithm module is "algorithm module A". Figure 2 As shown, the input port of algorithm module A is "color image to grayscale image - color image", and the output port is "image binarization - mask image".
[0090] The input port is configured based on a first mapping parameter, the first mapping parameter including a mapping name, Figure 2 In the mapping name of the input port, the mapping name is "Color Image to Grayscale Image - Color Image", and the mapping relationship is configured in the mapping name. Figure 2 The mapping relationship in the mapping name of the input port is that the input port is mapped to an operator unit having a color image input terminal and a processing logic name of "color image to grayscale image".
[0091] The output port is configured based on a second mapping parameter, the second mapping parameter includes a mapping name, Figure 2 In the output port mapping name, the mapping name is "Image Binarization - Mask Image", and the mapping relationship is configured in the mapping name. Figure 2 The mapping relationship in the mapping name of the output port in is that the output port is mapped to an operator unit having a mask image output terminal and a processing logic name of “image binarization”.
[0092] Figure 3 An operator unit sequence of one embodiment of the present disclosure is exemplarily shown.
[0093] Figure 3 The operator unit sequence shown includes only one input side operator unit and one output side operator unit. The processing logic name of the input side operator unit, i.e., the operator name, is "color image to grayscale image", and the processing logic name of the output side operator unit, i.e., the operator name, is "image binarization".
[0094] In some embodiments of the present disclosure, the input configuration parameters of the operator unit include the input data type and the input data name, the output configuration parameters of the operator unit include the output data type and the output data name, and the operator unit is configured with a processing logic name (operator name) that characterizes the function of the processing logic of the operator unit.
[0095] refer to Figure 3 The data type (i.e. parameter type) of the input configuration parameter of the input side operator unit is the color image type, and the data name (i.e. parameter name) is color image, that is, the input end can and can only input color image type data, and the data type of the output configuration parameter of the input side operator unit is the grayscale image type, and the data name is grayscale image, and the output end outputs a grayscale image.
[0096] The data type (also known as the parameter type) of the input configuration parameter of the output-side operator unit is a grayscale image type, and the data name (also known as the parameter name) is a grayscale image, that is, the input end can and can only input data of the grayscale image type. The data type of the output configuration parameter of the output-side operator unit is a mask image type, and the data name is a mask image, and the output end outputs a mask image.
[0097] Depend on Figure 3 It can be seen that data links are configured between the operator units of the operator unit sequence of the present disclosure, and the data links between the operator units are configured based on the data flow direction between the operator units.
[0098] Figure 4 The parameter configuration interface of the input port of the algorithm module (i.e., subgraph) of one embodiment of the present disclosure is exemplarily shown. Figure 4 The parameter configuration interface shown can be used to modify the mapping name of the first mapping parameter (ie, "input parameter"), add comments, and other operations.
[0099] In some embodiments of the present disclosure, you can click on, for example, Figure 2 Activate when the input port ("Color Image to Grayscale Image - Color Image") or output port ("Image Binarization - Mask Image") shown in Figure 4 The parameter configuration interface shown in the figure can switch between displaying the first mapping parameter (ie, "input parameter") and the second mapping parameter (ie, "output parameter").
[0100] Figure 5 It is a flowchart of a method for constructing a first mapping parameter according to an embodiment of the present disclosure.
[0101] refer to Figure 5 In some embodiments of the present disclosure, in the algorithm module configuration method of the present disclosure, generating a first mapping parameter based on the processing logic name and input terminal configuration parameter of the input side operator unit of the operator unit sequence includes:
[0102] S1041. Extract the processing logic name and input terminal configuration parameters of the input-side operator unit.
[0103] S1042. Construct a mapping name of the first mapping parameter (e.g., "color image to grayscale image - color image") based on the input data name and the processing logic name, use the input data type as the data type of the first mapping parameter, and construct a mapping relationship of the first mapping parameter based on the data type of the first mapping parameter (e.g., color image data) and the input data type (e.g., color image data).
[0104] Based on the above method, the present disclosure can realize automatic construction of the first mapping parameter.
[0105] Figure 6 It is a flowchart of a method for constructing a second mapping parameter according to an embodiment of the present disclosure.
[0106] refer to Figure 6 In some embodiments of the present disclosure, in the algorithm module configuration method of the present disclosure, generating the second mapping parameter based on the processing logic name and output terminal configuration parameter of the output side operator unit of the operator unit sequence includes:
[0107] S1041. Extract the processing logic name and output terminal configuration parameters of the output-side operator unit.
[0108] S1042. Construct a mapping name of the second mapping parameter (e.g., "image binarization - mask image") based on the output end data name and the processing logic name, use the output end data type as the data type of the second mapping parameter, and construct a mapping relationship of the second mapping parameter based on the data type of the second mapping parameter (e.g., mask image data) and the output end data type (e.g., mask image data).
[0109] Based on the above method, the present disclosure can realize automatic construction of the second mapping parameter.
[0110] In the algorithm module configuration method disclosed in the present invention, preferably, configuring the input node of the directed acyclic graph (DAG) based on the first mapping parameter includes:
[0111] The data type in the first mapping parameter is used as the input data type of the input node, the mapping name in the first mapping parameter is used as the input data name of the input node, and the mapping between the input node and the input end of the input side operator unit is established based on the mapping relationship in the first mapping parameter.
[0112] Preferably, configuring an output node of a directed acyclic graph (DAG) based on the second mapping parameter comprises:
[0113] The data type in the second mapping parameter is used as the output data type of the output node, the mapping name in the second mapping parameter is used as the output data name of the output node, and the mapping between the output node and the output end of the output side operator unit is established based on the mapping relationship in the second mapping parameter.
[0114] Based on the above method, the present disclosure can realize automatic configuration of the input nodes of the algorithm module based on the first mapping parameter, and realize automatic configuration of the output nodes of the algorithm module based on the second mapping parameter.
[0115] In some embodiments of the present disclosure, in the process of constructing an operator unit sequence of a certain algorithm module, multiple input-side operator units (or output-side operator units) with the same functions (i.e., the same processing logic) may be constructed. These input-side operator units (or output-side operator units) with the same functions can receive data of different operation scenarios (e.g., grasping scenarios), and different operation scenarios can be operation scenarios for grasping different types of target objects. The input-end (or output-end) data types of these input-side operator units (or output-side operator units) with the same functions are the same, but the data names are different, and multiple first mapping parameters (or multiple second mapping parameters) will be generated. Based on the multiple first mapping parameters, data of different scenarios can be received via the input node, or based on the multiple second mapping parameters, data can be output to different algorithm modules (subgraphs) via the output node.
[0116] The operator unit sequence disclosed in the present invention may include multiple input-side operator units with the same processing logic, the same input data type and different input data names, each input-side operator unit maps an input port of an input node, and each input port is configured based on a first mapping parameter.
[0117] Figure 7 The schematic diagram of the input nodes and output nodes of an algorithm module configured with multiple input ports and multiple output ports (before merging) is exemplarily shown.
[0118] Figure 8 Shows Figure 7 Parameter configuration interface of the algorithm module in .
[0119] refer to Figure 7 and Figure 8 , because three 3D precise matching operator units are configured in the operator unit sequence, the processing logic of these three operator units is the same, and the input data type is the same (both are scene point clouds), but the mapping names (i.e., input data names) of the three input ports of the algorithm module built based on these three operator units are different (scene point cloud 1, scene point cloud 2, scene point cloud 3); two 3D precise matching (unable to be directed) operator units are also configured in the operator unit sequence, the processing logic of these two operator units is the same, and the input data type is the same (both are scene point clouds), but the mapping names (i.e., input data names) of the two input ports of the algorithm module built based on these two operator units are different (unable to be directed - scene point cloud 1, unable to be directed - scene point cloud 2). Resulting in Figure 7 The algorithm module in the algorithm module has multiple input ports that only have different input data names, which makes the structure and interface of the algorithm module too complicated.
[0120] Fig. 9 A schematic flow chart of an algorithm module configuration method according to another embodiment of the present disclosure is shown, which includes an algorithm module optimization method (steps S106-S110).
[0121] To this end, in some preferred embodiments of the present disclosure, reference Fig. 9 , the algorithm module configuration method of the present disclosure further includes:
[0122] S106. Determine multiple input-side operator units whose inputs need to be merged based on the received operation instruction.
[0123] S108 . Performing consistency processing on the input-end data names of multiple input-side operator units that need to be merged and inputted, and generating a unified input-end data name.
[0124] S110. Merge multiple input ports corresponding to multiple input-side operator units that need to merge inputs (i.e., multiple input ports of the input node) into one input port, and generate a mapping name for the merged input port based on a unified input-end data name and a processing logic name of the input-side operator unit.
[0125] In some embodiments of the present disclosure, the operator unit sequence of the algorithm module of the present disclosure includes multiple output-side operator units with the same processing logic, the same output data type and different output data names, each output-side operator unit maps an output port of an output node, and each output port is configured based on a second mapping parameter.
[0126] In some implementations of the present disclosure, step S106 may further include: determining a plurality of output-side operator units that need to be combined for output based on the received operation instruction.
[0127] In some implementations of the present disclosure, step S108 may further include: performing uniform processing on the output-end data names of multiple output-side operator units that need to be combined and outputted, to generate a unified output-end data name.
[0128] In some embodiments of the present disclosure, step S110 may also include: merging multiple output ports corresponding to multiple output-side operator units that need to be merged into one output port (i.e., multiple output ports of the output node), and generating a mapping name for the merged output port based on a unified output port data name and a processing logic name of the output-side operator unit.
[0129] Fig.10 yes Figure 7 The schematic diagram of the algorithm module after the input nodes are merged. Fig.11 Shows Fig.10 Parameter configuration interface of the algorithm module in .
[0130] Based on the above algorithm module optimization method, the input nodes and / or output nodes of the algorithm module may be optimized, so that the structure of the input nodes and / or output nodes of the algorithm module is more concise.
[0131] refer to Figure 2 , Figure 7 , Fig.10 In a preferred embodiment of the present disclosure, step S104 of the algorithm module configuration method of the present disclosure further includes:
[0132] An enable switch is configured for the algorithm module based on the judgment logic (Bool), and the algorithm module can be executed based on the true value (true) or false value (false) of the judgment logic received by the enable switch.
[0133] In some embodiments of the present disclosure, when the enable switch receives a true value, the algorithm module can be executed, and when the enable switch receives a false value, the algorithm module cannot be executed.
[0134] In some other embodiments of the present disclosure, when the enable switch receives a false value, the algorithm module can be executed, and when the enable switch receives a true value, the algorithm module cannot be executed.
[0135] Based on the above description, the present disclosure implements effective configuration of algorithm modules (subgraphs) based on directed acyclic graphs.
[0136] Based on the algorithm module configured in any one of the embodiments of the present disclosure, the present disclosure further provides an algorithm execution component (ie, an algorithm diagram, also known as an algorithm device). Fig.12 A schematic structural block diagram of an algorithm execution component according to an embodiment of the present disclosure is shown.
[0137] refer to Fig.12 In one embodiment of the present disclosure, the algorithm execution component 2000 includes:
[0138] The target data output module 2002 (eg, "EpicPro input" module) processes the input data and outputs the target data.
[0139] The conditional judgment module 2004 is configured with conditional judgment logic. The conditional judgment module 2004 receives target data, and performs conditional judgment on the target data based on its own conditional judgment logic, and outputs a true value or a false value.
[0140] There are two algorithm modules 2006 , one algorithm module 2006 can be executed based on the true value output by the condition judgment module 2004 , and the other algorithm module 2006 can be executed based on the false value output by the condition judgment module 2004 .
[0141] The target data may include two-dimensional images and three-dimensional point clouds captured by the camera.
[0142] Among them, the conditional judgment logic described above can be target object category judgment logic, distance judgment logic between the target object and the camera, or target object orientation judgment logic (for example, judging whether the point cloud is the upper surface or the lower surface of the target object).
[0143] Those skilled in the art should understand that the two algorithm modules 2006 belong to two algorithm execution links respectively, and each algorithm execution link may include more algorithm modules.
[0144] For example Fig.13 In the two algorithm execution chains shown in FIG. 1 , the first algorithm execution chain includes a branch sub-chain. Figure 1 The second algorithm execution chain includes branch sub Figure 2 , in the branch Figure 1 , branch Figure 2 Afterwards, other algorithm modules may be connected to form a complete first algorithm execution link and a complete second algorithm execution link. Thus, the algorithm execution component 2000 of this embodiment can realize dual-scene switching.
[0145] The following is Fig.13 The algorithm execution component example shown illustrates dual scene switching.
[0146] In order to realize the branch logic control of dual scenes, the present disclosure uses multiple subgraphs (ie, multiple algorithm modules) to realize the separation of the grasping logic of different grasping objects in different scenes.
[0147] Fig.13The "general condition judgment" module in the embodiment of the condition judgment module 2004 of the present disclosure can judge the category of the target object, and then execute the branch sub-item based on the judgment result. Figure 1 or branch Figure 2 , for example, if the target object is a bag, then execute the branch sub Figure 1 , if the target object is a box, execute the branch sub Figure 2 , or judge the distance between the target object and the camera, and then execute the branch sub- Figure 1 or branch Figure 2 , or judge the direction of the target object, and then execute the branch sub- Figure 1 or branch Figure 2 .
[0148] In this implementation, a "general condition judgment" module is used to implement the control judgment logic (true value or false value) of the branch, and the operation control of different subgraphs is implemented through the enable switch of the branch subgraph.
[0149] In some embodiments of the present disclosure, the dual scenes described above may be the following two scenes: 1. A scene in which the target object is relatively close to the camera; 2. A scene in which the target object is relatively far from the camera.
[0150] When the conditional judgment module 2004 performs conditional judgment based on the distance between the target object and the camera (i.e., the target data), when the distance is greater than the threshold, a true value is output, so that the branch sub Figure 1 Can be executed, when the distance is not greater than the threshold, output a false value, so that the branch sub Figure 2 can be executed.
[0151] For the sake of simplicity, the present disclosure Fig.13 The example of the target data output module 2002 is not shown in the figure, which may be an "EpicPro input" module, which may be referred to in Chinese patent application CN202410319704.9. Figure 4 And related text description.
[0152] Fig.14 FIG. 1 is a schematic block diagram showing the structure of an algorithm execution component according to another embodiment of the present disclosure. Fig.14 In this embodiment, the algorithm execution component 3000 includes:
[0153] The target data output module 3002 (eg, "EpicPro input" module) processes the input data and outputs the target data.
[0154] Multiple conditional judgment modules 3004 are provided, each conditional judgment module 3004 is configured with a conditional judgment logic, the conditional judgment logics of each conditional judgment module 3004 are different from each other, each conditional judgment module 3004 receives target data, and performs conditional judgment on the target data based on its own conditional judgment logic, and outputs a true value or a false value.
[0155] There are multiple Boolean operation units 3006, each of which is configured with a Boolean judgment logic. The Boolean judgment logics of the Boolean operation units 3006 are different from each other. Each Boolean operation unit 3006 is configured with a first input terminal, a second input terminal and a Boolean output terminal. The first input terminal of each Boolean operation unit 3006 receives a true value output by a condition judgment module 3004, and the second input terminal receives a true value output by another condition judgment module 3004. Each Boolean operation unit 3006 performs Boolean judgment on the truth values received by the first input terminal and the second input terminal based on its own Boolean judgment logic, and outputs a true value or a false value via the Boolean output terminal.
[0156] Multiple algorithm modules 3008 , each algorithm module 3008 is connected to a Boolean output terminal of a Boolean operation unit 3006 , and can be executed based on a true value or a false value output by the Boolean output terminal of the Boolean operation unit 3006 .
[0157] Among them, at least one algorithm module 3008 is an algorithm module configured by the algorithm module configuration method of the present disclosure.
[0158] The present disclosure configures multiple condition judgment modules and cooperates with multiple Boolean operation units to achieve switching between multiple scenes and switching between more subgraphs (ie, algorithm modules).
[0159] Fig.14 Two conditional judgment modules, namely conditional judgment module 1 and conditional judgment module 2, three Boolean operation units, namely Boolean operation unit 1, Boolean operation unit 2 and Boolean operation unit 3, and four algorithm modules (ie, subgraphs) are exemplarily shown.
[0160] Fig.15 is an example of the algorithm execution component 3000 of this embodiment. For the sake of simplicity of illustration, Fig.15 The target data output module 3002 is not shown.
[0161] refer to Fig.14 and Fig.15 For example, the upper "general condition judgment" module judges the shape characteristics of the target object based on the first condition, and the lower "general condition judgment" module judges the color characteristics of the target object based on the second condition.
[0162] The truth value of the first condition is output to the first input terminals of the three Boolean operation units, and the truth value of the second condition is output to the second input terminals of the three Boolean operation units.
[0163] For Boolean operation unit 1, for example, when the first input terminal and the second input terminal both input true values, the Boolean output terminal outputs a Boolean true value, triggering the branch sub Figure 3 The enabling switch makes the branch Figure 3 can be executed.
[0164] For the Boolean operation unit 3, for example, when the first input terminal and the second input terminal both input true values, the Boolean output terminal outputs a Boolean true value, triggering the branch sub Figure 4 The enabling switch makes the branch Figure 4 can be executed.
[0165] For the Boolean operation unit 2, for example, when the first input terminal and the second input terminal both input true values, the Boolean output terminal outputs a Boolean true value, triggering the branch sub Figure 1 The enabling switch makes the branch Figure 1 can be executed, when only one of the first input terminal and the second input terminal receives a true value, its Boolean output terminal outputs a Boolean false value, triggering the branch sub Figure 2 Enable switch.
[0166] It can be seen that the algorithm execution component 3000 of this embodiment can realize the switching of three or more branch subgraphs based on the cooperation of multiple condition judgment modules and multiple Boolean operation units, and each branch subgraph can correspond to a grasping scene of a target object.
[0167] In some embodiments of the present disclosure, the algorithm execution component 2000 or the algorithm execution component 3000 described above in the present disclosure may be used as a subgraph of another algorithm execution component.
[0168] Fig.16 The complete execution chain of an algorithm execution component example of the present disclosure is shown.
[0169] refer to Fig.16 The target data output by the "EpicPro Input" module includes a 3D point cloud, a hand-eye calibration matrix, and a 3D ROI area. One 3D point cloud can be processed in sequence by the "Point Cloud ROI" module, the "Point Cloud Downsampling" module, the "Point Cloud Normal Estimation" module, and the "Point Cloud Edge Extraction (Unordered)" module to output a point cloud with normals and point cloud edges. The other 3D point cloud enters a conditional judgment module (i.e. Fig.16The "general condition judgment" module in the algorithm is used to judge whether the three-dimensional point cloud is the upper surface or the lower surface of the target object based on the feature information of the three-dimensional point cloud (such as the height of the point cloud). If it is the lower surface, a true value is output to the enable switch of the lower surface subgraph so that the lower surface subgraph can be executed. If it is the upper surface, the condition judgment module outputs a false value to the upper surface subgraph so that the upper surface subgraph can be executed, thereby realizing the switching of the subgraph branch, that is, realizing the switching of the execution link of the algorithm execution component.
[0170] The present disclosure also provides a readable storage medium, in which the algorithm execution component of any embodiment of the present disclosure is stored in the form of a computer program.
[0171] The present disclosure also provides a computer program product, which is an algorithm execution component of any embodiment of the present disclosure.
[0172] The present disclosure also provides a vision-based robot system, comprising:
[0173] Camera, the camera is used to obtain a two-dimensional image and a three-dimensional point cloud of the target object;
[0174] The algorithm execution component of any embodiment of the present disclosure processes the two-dimensional image and the three-dimensional point cloud acquired by the camera to obtain the position and posture information of the target object;
[0175] A robot operates a target object (eg, grasping operation) based on the position information of the target object.
[0176] The robot disclosed herein may be a grasping robot with a multi-degree-of-freedom mechanical arm. The target object may be a box, a bag, or the like.
[0177] The method of the present disclosure may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented by software, it may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instruction is loaded and executed, the process or function of the present disclosure is executed in whole or in part.
[0178] The computer program or instructions may be stored in a readable storage medium or transmitted from one readable storage medium to another readable storage medium, for example, the computer program or instructions may be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired or wireless means. The readable storage medium may be any available medium that can be accessed or a data storage device such as a server, data center, etc. that integrates one or more available media. The available medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; it may also be an optical medium, such as a digital video disk; it may also be a semiconductor medium, such as a solid state drive. The computer readable storage medium may be a volatile or non-volatile storage medium, or may include both volatile and non-volatile types of storage media.
[0179] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0180] The present disclosure is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0181] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0182] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0183] In the description of this specification, the description with reference to the terms "one embodiment / method", "some embodiments / methods", "example", "specific example", or "some examples" etc. means that the specific features, structures, or characteristics described in conjunction with the embodiment / method or example are included in at least one embodiment / method or example of the present disclosure. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment / method or example. Moreover, the specific features, structures, or characteristics described may be combined in any one or more embodiments / methods or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments / methods or examples described in this specification and features of different embodiments / methods or examples, unless they are contradictory.
[0184] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present disclosure, "plurality" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.
[0185] Those skilled in the art should understand that the above embodiments are only for the purpose of clearly illustrating the present disclosure, and are not intended to limit the scope of the present disclosure. For those skilled in the art, other changes or modifications may be made based on the above disclosure, and these changes or modifications are still within the scope of the present disclosure.
Claims
1. An algorithm module configuration method, characterized in that: include: Based on the multiple atomic nodes of the directed acyclic graph and the directed edges between the atomic nodes, multiple operator units and the data flow between the operator units are configured to generate an operator unit sequence of the algorithm module, wherein the operator unit includes an input end, an output end, and a processing logic configured between the input end and the output end; as well as Generate a first mapping parameter based on the processing logic name and input terminal configuration parameter of the input-side operator unit of the operator unit sequence, configure the input node of the directed acyclic graph based on the first mapping parameter, and the input node serves as the input port of the algorithm module, so that the data received by the input port can be mapped to the input terminal of the input-side operator unit based on the first mapping parameter; A second mapping parameter is generated based on the processing logic name and output end configuration parameter of the output-side operator unit of the operator unit sequence, and the output node of the directed acyclic graph is configured based on the second mapping parameter. The output node serves as the output port of the algorithm module, so that the data output through the output end of the output-side operator unit can be mapped to the output port of the algorithm module based on the second mapping parameter.
2. The algorithm module configuration method according to claim 1, characterized in that: Data links are configured between the operator units of the operator unit sequence, and the data links between the operator units are configured based on the data flow direction between the operator units.
3. The algorithm module configuration method according to claim 1, characterized in that: The atomic node is an inseparable node in the directed acyclic graph.
4. The algorithm module configuration method according to any one of claims 1 to 3, characterized in that: The input configuration parameters of the operator unit include the input data type and the input data name, the output configuration parameters of the operator unit include the output data type and the output data name, and the operator unit is configured with a processing logic name that characterizes the function of the processing logic of the operator unit; The first mapping parameter is generated based on the processing logic name and input terminal configuration parameter of the input side operator unit of the operator unit sequence, including: Extracting the processing logic name and input terminal configuration parameters of the input-side operator unit; and Building a mapping name of the first mapping parameter based on the input end data name and the processing logic name, taking the input end data type as the data type of the first mapping parameter, and building a mapping relationship of the first mapping parameter based on the data type of the first mapping parameter and the input end data type; Optionally, the generating of the second mapping parameter based on the processing logic name and output terminal configuration parameter of the output-side operator unit of the operator unit sequence includes: Extracting the processing logic name and output terminal configuration parameters of the output side operator unit; and Building a mapping name of the second mapping parameter based on the output end data name and the processing logic name, taking the output end data type as the data type of the second mapping parameter, and building a mapping relationship of the second mapping parameter based on the data type of the second mapping parameter and the output end data type; Optionally, configuring the input node of the directed acyclic graph based on the first mapping parameter includes: Using the data type in the first mapping parameter as the input data type of the input node, using the mapping name in the first mapping parameter as the input data name of the input node, and establishing a mapping between the input node and the input end of the input-side operator unit based on the mapping relationship in the first mapping parameter; Optionally, configuring the output node of the directed acyclic graph based on the second mapping parameter includes: Using the data type in the second mapping parameter as the output data type of the output node, using the mapping name in the second mapping parameter as the output data name of the output node, and establishing a mapping between the output node and the output terminal of the output-side operator unit based on the mapping relationship in the second mapping parameter; Optionally, the operator unit sequence includes a plurality of input-side operator units with the same processing logic, the same input-end data type and different input-end data names, each input-side operator unit maps an input port of the input node, and each input port is configured based on a first mapping parameter; The configuration method further includes: Determine multiple input-side operator units whose inputs need to be merged; Performing uniform processing on the input-end data names of the multiple input-side operator units that need to be merged and inputted, and generating a unified input-end data name; and Merge the multiple input ports corresponding to the multiple input-side operator units that need to merge inputs into one input port, and generate a mapping name of the merged input port based on the unified input end data name and the processing logic name of the input-side operator unit; Optionally, the operator unit sequence includes a plurality of output-side operator units with the same processing logic, the same output-end data type and different output-end data names, each output-side operator unit maps an output port of the output node, and each output port is configured based on a second mapping parameter; The configuration method further includes: Determine multiple output-side operator units that need to be combined for output; Performing uniform processing on the output-end data names of the multiple output-side operator units that need to be combined and outputted, and generating a unified output-end data name; and Merge the multiple output ports corresponding to the multiple output-side operator units that need to be merged into one output port, and generate a mapping name for the merged output port based on the unified output port data name and the processing logic name of the output-side operator unit; Optionally, it also includes: An enabling switch is configured for the algorithm module based on the judgment logic, and the algorithm module can be executed based on a true value or a false value of the judgment logic received by the enabling switch.
5. An algorithm module, characterized in that: The algorithm module is an algorithm module configured based on the configuration method described in any one of claims 1 to 4.
6. An algorithm execution component, characterized in that: include: A target data output module, which processes the input data and outputs the target data; A conditional judgment module, wherein the conditional judgment module is configured with conditional judgment logic, receives the target data, and performs conditional judgment on the target data based on its own conditional judgment logic, and outputs a true value or a false value; as well as Two algorithm modules, one algorithm module can be executed based on the true value output by the condition judgment module, and the other algorithm module can be executed based on the false value output by the condition judgment module; Wherein, at least one of the algorithm modules is the algorithm module described in claim 5.
7. An algorithm execution component, characterized in that: include: A target data output module, which processes the input data and outputs the target data; A plurality of conditional judgment modules, each of which is configured with a conditional judgment logic, the conditional judgment logics of the conditional judgment modules are different from each other, each of the conditional judgment modules receives the target data, and performs conditional judgment on the target data based on its own conditional judgment logic, and outputs a true value or a false value; A plurality of Boolean operation units, each of which is configured with a Boolean judgment logic, the Boolean judgment logics of the Boolean operation units are different from each other, each of which is configured with a first input terminal, a second input terminal and a Boolean output terminal, the first input terminal of each Boolean operation unit receives a true value output by a condition judgment module, the second input terminal receives a true value output by another condition judgment module, each Boolean operation unit performs a Boolean judgment on the true values received by the first input terminal and the second input terminal based on its own Boolean judgment logic, and outputs a true value or a false value via the Boolean output terminal; as well as a plurality of algorithm modules, each algorithm module being connected to a Boolean output terminal of a Boolean operation unit and capable of being executed based on a true value or a false value output by the Boolean output terminal of the Boolean operation unit; Wherein, at least one of the algorithm modules is the algorithm module according to claim 5; Optionally, two algorithm modules among the plurality of algorithm modules are connected to the Boolean output terminal of the same Boolean operation unit, one of the two algorithm modules can be executed based on a true value output by the Boolean operation unit, and the other algorithm module can be executed based on a false value output by the Boolean operation unit; Optionally, the target data includes a two-dimensional image and a three-dimensional point cloud acquired by a camera; Optionally, the conditional judgment logic is a target object category judgment logic, a distance judgment logic between the target object and the camera, or a target object direction judgment logic.
8. A readable storage medium, characterized in that: The algorithm execution component of claim 6 or 7 is stored in the readable storage medium in the form of a computer program.
9. A computer program product, characterized in that The computer program product is the algorithm execution component according to claim 6 or 7.
10. A vision-based robotic system, characterized in that: include: A camera, wherein the camera is used to obtain a two-dimensional image and a three-dimensional point cloud of a target object; The algorithm execution component of claim 6 or 7, wherein the algorithm execution component processes the two-dimensional image and three-dimensional point cloud acquired by the camera to obtain the position and posture information of the target object; as well as A robot operates the target object based on the position information of the target object.
