A model-driven path planning method for surface data acquisition of engine castings

Abstract

This invention discloses a model-driven path planning method for collecting engine casting surface data. The method uses a laser scanner to acquire depth image information of the engine casting surface; establishes an engine casting surface obstacle model based on this depth image information; employs a minimum cost spanning tree path planning algorithm to plan a camera's movement path along the engine casting surface to acquire image data based on the obstacle model; and uses a robotic arm to drive the camera along the movement path to complete the acquisition of engine casting surface image data. This method boasts high acquisition efficiency and stable path planning, significantly improving the stability of the engine casting surface data acquisition process and ensuring the quality of the acquired data.

Description

Technical Field

[0001] The present invention belongs to the technical field of engine castings, and in particular relates to a model-driven method for planning a path for collecting surface data of an engine casting. Background Art

[0002] Castings are metal shaped objects obtained using various casting methods, such as pouring, injection, suction or other casting methods into a pre-prepared casting model after smelting liquid metal, and then polishing and other processing methods to obtain metal objects with a certain shape, size and performance.

[0003] Automobile engine castings are one of the most difficult types of castings due to their complex structure, high dimensional accuracy requirements, and stringent requirements on casting surface quality and strength.

[0004] Over the years, the casting process for engine castings has made significant progress. However, due to factors such as production processes and operator handling, surface defects inevitably appear during the casting process. These defects can pose serious safety hazards during subsequent use of the engine castings. Therefore, it is necessary to inspect the surface of castings for defects before they leave the factory.

[0005] Before conducting surface inspections on engine castings, obtaining high-quality, stable and efficient surface images is a critical step to ensuring smooth progress. In current production practices, surface data collection for engine castings primarily relies on manual photography. However, this method is inefficient and unreliable, time-consuming, and labor-intensive, making it difficult to meet the demands of efficient production. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a model-driven engine casting surface data acquisition path planning method in response to the above-mentioned deficiencies in the prior art.

[0007] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0008] A model-driven method for engine casting surface data acquisition path planning includes:

[0009] Step 1: Use a laser scanner to obtain depth image information of the engine casting surface;

[0010] Step 2: establishing an engine casting surface obstacle model based on the engine casting surface depth image information;

[0011] Step 3: Using a minimum cost spanning tree path planning algorithm, the camera is planned to acquire image data on the surface of the engine casting based on the obstacle model.

[0012] Step 4: Use the robotic arm to drive the camera to complete the image data acquisition of the engine casting surface along the motion path.

[0013] To optimize the above technical solutions, specific measures taken also include:

[0014] The engine casting surface depth image information obtained in step 1 above is in the format of (x, y, z)×H×W;

[0015] Where H and W represent the sampling length and width of the engine casting surface, that is, the number of sampling points on the length and width of the casting; x and y represent the horizontal and vertical coordinates of the sampling points; and z represents the height coordinate of the sampling point.

[0016] The above height coordinates are the values obtained through the unified processing, specifically: take the maximum value of the height coordinate z in the depth image information max To reference the origin, modify all other height coordinate values to match z max The relative difference between , which is non-negative.

[0017] The engine casting surface obstacle model established in step 2 above is S ob =[i,x ob ,y ob ,l x ,l y ,μ i ];

[0018] Among them, i represents the number of the current obstacle block, x ob ,y ob Indicates the horizontal calibration position coordinates of each obstacle block;

[0019] l x ,l y Represents the horizontal width and vertical length of the obstacle block, which are calculated based on the horizontal coordinates x and y of the sampling point respectively;

[0020] μ i Represents the height barrier factor.

[0021] The above-mentioned height obstacle factor is calculated by the following formula:

[0022]

[0023] Among them, max(z) i Indicates the maximum value of the height coordinate within the obstacle block range, min(z) i Indicates the minimum value of the height coordinate within the obstacle block range.

[0024] Step 3 above includes the following steps:

[0025] Step 3.1: Calculate the weight value ω between the adjacent i-th and j-th obstacle blocks based on the height obstacle factor of each obstacle block. i,j =|μ i -μ j |, the weight has no negative value;

[0026] Step 3.2: With the weight values as the weight edges of the adjacent obstacle blocks, the entire engine casting surface obstacle model is abstracted into a weighted undirected graph A, and the minimum cost tree set {γ 0→n}, n represents the number of minimum cost trees;

[0027] Step 3.3: For each minimum cost tree, select the initial coordinates of the camera model, that is, the point representing the first obstacle block as the starting point of the path, traverse the nodes representing the obstacle blocks in the entire tree, and calculate the traversal weight. Get the traversal weight set {W n};

[0028] Among them, OP represents the number of traversal operations, ω k Indicates the weight of the operation edge in the minimum cost tree corresponding to the k-th traversal operation;

[0029] Step 3.4, filter and obtain the traversal weight set {W n The traversal path corresponding to the minimum value in} is the action path output by the path planning algorithm of the minimum cost spanning tree, which is used to plan the camera based on the obstacle model to obtain image data on the surface of the engine casting.

[0030] The undirected graph A described in step 3.2 above is:

[0031]

[0032] The initial coordinates of the camera model in step 3.3 are (x c ,y c ,z c );

[0033] Among them, x c ,y c z is the horizontal and vertical coordinates of the upper left corner of the engine casting surface, c is the maximum value of the height coordinate z of the sampling point in the depth image information of the engine casting surface max .

[0034] The present invention has the following beneficial effects:

[0035] The present invention uses a laser scanner to acquire depth image information of the engine casting surface; establishes an engine casting surface obstacle model based on this depth image information; employs a minimum cost spanning tree path planning algorithm to plan a camera's movement path along the engine casting surface to acquire image data based on the obstacle model; and uses a robotic arm to drive the camera along the movement path to complete the acquisition of engine casting surface image data. This system boasts high acquisition efficiency and stable path planning, significantly improving the stability of the engine casting surface data acquisition process and ensuring the quality of the acquired data. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Schematic diagram of the flow of the model-driven engine casting surface data acquisition path planning method of the present invention;

[0037] Figure 2 Schematic diagram of the division of obstacle blocks in the engine casting surface obstacle model of the present invention;

[0038] Figure 3 This is a schematic diagram of the engine casting surface obstacle model in the present invention being abstracted as an undirected graph. DETAILED DESCRIPTION

[0039] The embodiments of the present invention are described in further detail below with reference to the accompanying drawings.

[0040] See also Figure 1 , a model-driven engine casting surface data acquisition path planning method, comprising:

[0041] Step 1: Use a laser scanner to obtain depth image information of the engine casting surface;

[0042] The engine casting surface depth image information obtained in step 1 is in the format of (x, y, z)×H×W;

[0043] Where H and W represent the sampling length and width of the engine casting surface, that is, the number of sampling points on the length and width of the casting; x and y represent the horizontal and vertical coordinates of the sampling points; and z represents the height coordinate of the sampling point.

[0044] The height coordinate is the value obtained through the unified processing, specifically: take the maximum value of the height coordinate z in the depth image information max To reference the origin, modify all other height coordinate values to match z max The relative difference between , which is non-negative.

[0045] Step 2: establishing an engine casting surface obstacle model based on the engine casting surface depth image information;

[0046] The engine casting surface obstacle model established in step 2 is S ob=[i,x ob ,y ob ,l x ,l y ,μ i ];

[0047] Among them, i represents the number of the current obstacle block, x ob ,y ob Indicates the horizontal calibration position coordinates of each obstacle block;

[0048] l x ,l y Represents the horizontal width and vertical length of the obstacle block, which are calculated based on the horizontal coordinates x and y of the sampling point respectively;

[0049] μ i represents the height barrier factor, which is calculated as follows:

[0050]

[0051] Among them, max(z) i Indicates the maximum value of the height coordinate within the obstacle block range, min(z) i Indicates the minimum value of the height coordinate within the obstacle block range.

[0052] The schematic diagram of the obstacle block division in the engine casting surface obstacle model is as follows: Figure 2 shown.

[0053] Step 3, using a minimum cost spanning tree path planning algorithm to plan the camera's action path for acquiring image data on the engine casting surface based on the obstacle model, includes the following steps:

[0054] Step 3.1: Calculate the weight value ω between the adjacent i-th and j-th obstacle blocks based on the height obstacle factor of each obstacle block. i,j =|μ i -μ j |, the weight has no negative value;

[0055] Step 3.2, such as Figure 3 As shown in the figure, the weight value is the weight edge of the adjacent obstacle block, the entire engine casting surface obstacle model is abstracted into a weighted undirected graph A, and the minimum cost tree set {γ 0→n}, n represents the number of minimum cost trees;

[0056] The undirected graph A is an N×N weight matrix, and the value of each point (i, j) in the weight matrix represents the weight value ω between the adjacent i-th and j-th obstacle blocks i,j :

[0057]

[0058] Step 3.3: For each minimum cost tree, select the initial coordinates of the camera model, that is, the point representing the first obstacle block as the starting point of the path, traverse the nodes representing the obstacle blocks in the entire tree, and calculate the traversal weight. Get the traversal weight set {W n};

[0059] Among them, OP represents the number of traversal operations, ω k Indicates the weight of the operation edge in the minimum cost tree corresponding to the k-th traversal operation;

[0060] Step 3.4, filter and obtain the traversal weight set {W n The traversal path corresponding to the minimum value in} is the action path output by the path planning algorithm of the minimum cost spanning tree, which is used to plan the camera based on the obstacle model to obtain image data on the surface of the engine casting.

[0061] The initial coordinates of the camera model are (x c ,y c ,z c );

[0062] Among them, x c ,y c z is the horizontal and vertical coordinates of the upper left corner of the engine casting surface, c is the maximum value of the height coordinate z of the sampling point in the depth image information of the engine casting surface max .

[0063] Step 4: Use the robotic arm to drive the camera to complete the image data acquisition of the engine casting surface along the motion path.

[0064] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions based on the principles of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A model-driven method for surface data acquisition path planning of engine castings, characterized in that: include: Step 1: Use a laser scanner to obtain depth image information of the engine casting surface; Step 2: establishing an engine casting surface obstacle model based on the engine casting surface depth image information; Step 3: Using a minimum cost spanning tree path planning algorithm, the camera is planned to acquire image data on the surface of the engine casting based on the obstacle model. Step 4: Use the robotic arm to drive the camera to complete the image data acquisition of the engine casting surface along the motion path; The engine casting surface depth image information obtained in step 1 is in the format of (x, y, z)×H×W; Where H and W represent the sampling length and width of the engine casting surface, that is, the number of sampling points on the length and width of the casting; x and y represent the horizontal and vertical coordinates of the sampling points; z represents the height coordinate of the sampling point; The engine casting surface obstacle model established in step 2 is S ob =[i,x ob ,y ob ,l x ,l y ,μ i ]; Among them, i represents the number of the current obstacle block, x ob ,y ob Indicates the horizontal calibration position coordinates of each obstacle block; l x ,l y represents the horizontal width and vertical length of the obstacle block, which are calculated based on the horizontal coordinates x and y of the sampling point respectively; μ i represents the height barrier factor; the height barrier factor is calculated by the following formula: Among them, max(z) i Indicates the maximum value of the height coordinate within the obstacle block range, min(z) i Indicates the minimum value of the height coordinate within the obstacle block range.

2. The model-driven engine casting surface data acquisition path planning method according to claim 1, characterized in that: The height coordinate is a value obtained through a unified process, specifically: taking the maximum value of the height coordinate z in the depth image information max To reference the origin, modify all other height coordinate values to match z max The relative difference is non-negative.

3. The model-driven engine casting surface data acquisition path planning method according to claim 1, characterized in that: Step 3 includes the following steps: Step 3.1: Calculate the weight value ω between the adjacent i-th and j-th obstacle blocks based on the height obstacle factor of each obstacle block. i,j =|μ i -μ j |, the weight has no negative value; Step 3.2: With the weight values as the weight edges of the adjacent obstacle blocks, the entire engine casting surface obstacle model is abstracted into a weighted undirected graph A, and the minimum cost tree set {γ 0→n }, n represents the number of minimum cost trees; Step 3.3: For each minimum cost tree, select the initial coordinates of the camera model, that is, the point representing the first obstacle block as the starting point of the path, traverse the nodes representing the obstacle blocks in the entire tree, and calculate the traversal weight. Get the traversal weight set {W n }; Among them, OP represents the number of traversal operations, ω k Indicates the weight of the operation edge in the minimum cost tree corresponding to the k-th traversal operation; Step 3.4, filter and obtain the traversal weight set {W n The traversal path corresponding to the minimum value in} is the action path output by the path planning algorithm of the minimum cost spanning tree, which is used to plan the camera based on the obstacle model to obtain image data on the surface of the engine casting.

4. The model-driven engine casting surface data acquisition path planning method according to claim 3, characterized in that: The undirected graph A described in step 3.2 is:

5. The model-driven engine casting surface data acquisition path planning method according to claim 3, characterized in that: The initial coordinates of the camera model in step 3.3 are (x c ,y c ,z c ); Among them, x c ,y c z is the horizontal and vertical coordinates of the upper left corner of the engine casting surface, c is the maximum value of the height coordinate z of the sampling point in the depth image information of the engine casting surface max .

Images