A new two-dimensional and three-dimensional data fusion air conditioner evaporator gluing path obtaining algorithm

Abstract

The application relates to an air conditioner evaporator gluing path obtaining method, discloses a novel two-dimensional and three-dimensional data fusion air conditioner evaporator gluing path obtaining algorithm, specifically, first, a two-dimensional camera and a three-dimensional camera are used to respectively shoot a brightness image imgA and a depth image imgB of an air conditioner evaporator; then, the depth image imgB and the brightness image imgA are respectively processed to obtain an interested image imgC, then the interested image imgC is subjected to connected region marking and sorting to obtain a target connected region, finally, a least square fitting method is used to fit and obtain the coordinates of gluing points, and the air conditioner evaporator gluing path is obtained. The application can obtain a complete gluing path, has high guiding precision, and can effectively guarantee the whole accurate gluing of the evaporator.

Description

TECHNICAL FIELD

[0001] The application relates to an air conditioner evaporator gluing path obtaining method, in particular to a novel air conditioner evaporator gluing path obtaining algorithm based on two-dimensional and three-dimensional data fusion. BACKGROUND

[0002] As a heat exchange carrier of an air conditioner, an evaporator plays a decisive role in the energy efficiency of the air conditioner. In order to ensure the sealing property of the evaporator, the evaporator needs to be glued during assembly. Traditional methods have the problems of slow efficiency and difficulty in meeting the requirements of peak demand in the peak season. In recent years, visual guidance of a mechanical hand for gluing has also been adopted. In order to improve the guidance accuracy, a two-dimensional camera and a three-dimensional camera are usually combined. However, the traditional two-dimensional camera and three-dimensional camera fusion method has the problem of low data integrity and difficulty in obtaining a complete gluing path. SUMMARY

[0003] The application provides a novel air conditioner evaporator gluing path obtaining algorithm based on two-dimensional and three-dimensional data fusion.

[0004] In order to solve the above technical problems, the application solves the problems through the following technical scheme:

[0005] A novel air conditioner evaporator gluing path obtaining algorithm based on two-dimensional and three-dimensional data fusion comprises the following steps:

[0006] Step S1, a two-dimensional camera and a three-dimensional camera are installed on an air conditioner evaporator gluing conveying belt, and a brightness image imgA and a depth image imgB of the air conditioner evaporator are respectively shot by the two-dimensional camera and the three-dimensional camera;

[0007] Step S2, binaryzation processing is performed on the depth image imgB to complete segmentation, and a binaryzation segmented image imgB1 is obtained;

[0008] Step S3, eight-connected region marking screening is performed on the binaryzation segmented image to obtain a binary image imgB2 after initial screening;

[0009] Step S4, general plastic parts contained in the screened binary image imgB2 are coarsely positioned, and an image imgB3 after coarse positioning is obtained;

[0010] Step S5, Gamma correction is performed on the brightness image imgA in step S1 to obtain a corrected image imgA1;

[0011] Step S6, binaryzation processing is performed on the corrected image imgA1, and then eight-connected region marking screening is used to obtain an image imgA2,

[0012] Step S8, subtract the image imgB3 after the coarse positioning in step S4 from the image imgA2 described in step 7 to obtain an image of interest imgC;

[0013] Step S9, label the connected regions of the image of interest imgC, count the geometric centers of each connected region, and sort the geometric centers according to the y coordinates from large to small;

[0014] Step S10, calculate the Euclidean distance between each connected region in step S9 and the connected region labeled 1, and collect the distances as {d1, d2,..., dn}, and take the connected region with the smallest distance to the preset value T as the target connected region; n

[0015] Step S11, for the connected region labeled 1, use the least square fitting method to fit a straight line y = b + a*x, wherein In the formula, x i and y i are the pixel values of the two bucket regions i, and the intersection of the straight line y = a*x + b and the leftmost side of the binary image imgB2 is P1, and the coordinates of the point P2 on the straight line y = a*x + b, which is right of P1 and has a distance D from P1.

[0016] Step S12, for the connected region labeled 3, use the least square fitting method to fit a straight line s = k*u + v. Record the intersection of the straight line s = k*u + v and the leftmost side of the binary image imgB2 as Q1, and the coordinates of the point Q2 on the straight line s = k*u + v, which is right of Q1 and has a distance D from Q1.

[0017] Step S13, obtain the coordinates of the glue application point: PT1 = w*P1 + (1-w)*Q1, PT2 = w*P2 + (1-w)*Q2, and then obtain the glue application path of the air conditioner evaporator.

[0018] By respectively processing the brightness image imgA and the depth image imgB captured by the two-dimensional camera and the three-dimensional camera, and obtaining the image of interest imgC according to the processed images, then continuing to process the image of interest imgC to obtain the target connected region, and then obtaining the coordinates of the glue application point according to a certain algorithm, and then obtaining the glue application path, the glue application path data obtained in this way has high integrity, and the data is obtained according to the actual shape of the air conditioner evaporator, which can effectively improve the visual guidance accuracy, and then the finally obtained glue application path has high precision and integrity, ensuring the accuracy and completeness of the glue application process, and effectively improving the product quality.

[0019] As a preferred, the two-dimensional camera and the three-dimensional camera in step S1 both adopt an external trigger mode, and the air conditioner evaporator triggers the photographing when passing through the flow line photographing station.​

[0020] As preferred, the binarization processing in step S2 and step S6 both adopt fixed threshold value to binarize the depth image, the fixed threshold value is set as 128, and the pixel value of all pixel points in the depth image imgB is compared one by one during binarization, the pixel value of the pixel point is greater than the fixed threshold value, then the pixel point is assigned value 1, and the pixel value of the pixel point is less than the fixed threshold value, then the pixel point is assigned value 0.

[0021] As preferred, the eight-connected region marking screening in step S3 and step S6 both remove the connected region with length less than 300.

[0022] As preferred, step S4 specifically includes the following steps:

[0023] Step S41, morphological erosion operation is performed on the screened binary image imgB2 to obtain the eroded binary image, and the formula is In the formula, S is the screened binary image imgB2, X is the eroded image, and x is the current position;

[0024] Step S42, morphological dilation operation is performed on the eroded binary image to obtain the dilated binary image, and the formula is In the formula, S is the eroded image in step S41, X is the dilated image, and x is the current position;

[0025] Step S43, connected region screening height operation is performed to screen out the height closest to the fixed height of the evaporator plastic part, and obtain the coarsely positioned image imgB3.

[0026] As preferred, the erosion window width in step S41 is 5, and the erosion window height is 51.

[0027] As preferred, the dilation window width in step S42 is 10, and the height window is 251.

[0028] As preferred, 80 pixel connected regions are selected as the region of the air conditioner evaporator plastic part in step S43. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 is the luminance image imgA taken by the two-dimensional camera in embodiment 1 of the present application.

[0030] Figure 2 is the depth image imgB taken by the three-dimensional camera in embodiment 1 of the present application.

[0031] Figure 3 is the binarized segmented image imgB1 in embodiment 1 of the present application.

[0032] Figure 4is the binary image imgB2 after the preliminary screening in the embodiment 1 of the present application.

[0033] Figure 5 is the image imgB3 after the rough positioning in the embodiment 1 of the present application.

[0034] Figure 6 is the image imgA1 after the correction in the embodiment 1 of the present application.

[0035] Figure 7 is the image imgA2 obtained by using the eight-connected region marking screening in the embodiment 1 of the present application.

[0036] Figure 8 is the image imgC of interest in the embodiment 1 of the present application.

[0037] Figure 9 is the image after the step S9 sorting in the embodiment 1 of the present application.

[0038] Figure 10 is the target connected region image in the step S10 in the embodiment 1 of the present application.

[0039] Figure 11 is the image of the gluing path in the step S13 in the embodiment 1 of the present application. DETAILED DESCRIPTION

[0040] The present application will be further described in detail below in combination with the drawings and embodiments.

[0041] Embodiment 1

[0042] A novel air conditioner evaporator gluing path obtaining algorithm of two-dimensional and three-dimensional data fusion includes the following steps:

[0043] Step S1, install a two-dimensional camera and a three-dimensional camera on an air conditioner evaporator gluing conveying belt, and shoot a brightness image imgA and a depth image imgB of the air conditioner evaporator through the two-dimensional camera and the three-dimensional camera respectively, as shown in Figure 1 、 Figure 2 In this embodiment, the two-dimensional camera and the three-dimensional camera are both in an external triggering mode, and when the air conditioner evaporator passes through the flow line shooting station, the two-dimensional camera and the three-dimensional camera are triggered to shoot the brightness image imgA and the depth image imgB respectively;

[0044] Step S2, complete the segmentation by performing a binaryzation processing on the depth image imgB, to obtain a binaryzation segmentation image imgB1, as shown in Figure 3As shown, in this embodiment, the binarization process uses a fixed threshold to perform binarization segmentation on the depth image. The fixed threshold is set to 128. During binarization, the pixel values ​​of all pixels in the depth image imgB are compared one by one. If the pixel value of a pixel is greater than the fixed threshold, the pixel is assigned a value of 1. If the pixel value of a pixel is less than the fixed threshold, the pixel is assigned a value of 0.

[0045] Step S3: Perform octagonal region labeling and filtering on the binarized segmented image, removing connected regions with a length less than 300, to obtain the initially filtered binary image imgB2, as shown. Figure 4 As shown;

[0046] Step S4: Perform coarse localization on the common plastic parts contained in the filtered binary image imgB2 to obtain the coarsely localized image imgB3, as shown. Figure 5 As shown, the specific steps include:

[0047] Step S41: Perform morphological erosion on the filtered binary image imgB2 to obtain the eroded binary image. The formula is as follows: In the formula, S is the filtered binary image imgB2, X is the eroded image, and x is the current position; in this embodiment, the erosion window width is 5 and the erosion window height is 51.

[0048] Step S42: Perform morphological dilation on the eroded binary image to obtain a dilated binary image, using the following formula: In the formula, S is the eroded image in step S41, X is the dilated image, and x is the current position; in this embodiment, the dilation window width is selected as 10 and the height window as 251.

[0049] Step S43: Perform connected region filtering height operation to filter out the region whose height is closest to the fixed height of the evaporator plastic part. In this embodiment, a connected region of 80 pixels is selected as the region where the air conditioner evaporator plastic part is located, and the image imgB3 after coarse positioning is obtained.

[0050] Step S5: Perform Gamma correction on the brightness image imgA from step S1. The correction formula is imgAGm = imgA 1 / gamma , gamma As the correction coefficient, in this embodiment, a correction coefficient of 3 is selected to obtain the corrected image imgA1, as shown. Figure 6 As shown;

[0051] Step S6: Binarize the corrected image imgA1. In this embodiment, a fixed threshold of 128 is used for binarization. Then, octal connected region labeling is used to filter out regions with a length greater than 300, resulting in image imgA2, as shown. Figure 7 As shown

[0052] Step S8, subtract the image imgB3 after the coarse positioning in step S4 from the image imgA2 described in step 7 to obtain the image imgC of interest, that is, imgC=imgA2-imgB3, as shown in Figure 8 .

[0053] Step S9, label the connected regions of the image imgC of interest, count the geometric centers of each connected region, and sort the geometric centers according to the y coordinates from large to small, as shown in Figure 9 .

[0054] Step S10, calculate the Euclidean distance between each connected region in step S9 and the connected region labeled 1, and collect the distances as {d1, d2,..., d n}, and take the connected region with the smallest distance to the preset value T=60 as the target connected region, as shown in Figure 10 .

[0055] Step S11, for the connected region labeled 1, use the least square fitting method to fit a straight line y=b+a*x, where In the formula, x i and y i are the pixel values of the two bucket regions i, and the intersection point of the straight line y=a*x+b and the leftmost side of the binary image imgB2 is P1, the point coordinates of the right end of P1 on the straight line y=a*x+b and the distance D=1450 from P1 are P2.

[0056] Step S12, for the connected region labeled 3, use the least square fitting method to fit a straight line s=k*u+v. The intersection point of the straight line s=k*u+v and the leftmost side of the binary image imgB2 is Q1, and the point coordinates of the right end of Q1 on the straight line s=k*u+v and the distance D=1450 from Q1 are Q2.

[0057] Step S13, obtain the coordinates of the glue application point: PT1=w*P1+(1-w)*Q1, PT2=w*P2+(1-w)*Q2, and take w=0.25 in this embodiment to obtain the glue application path of the air conditioner evaporator, as shown in Figure 11 .

[0058] The glue application path of the air conditioner evaporator obtained by the glue application path obtaining algorithm in this embodiment has complete data, can obtain a complete glue application path, and can be based on the actual shape of the air conditioner evaporator, thereby having high guiding accuracy and effectively ensuring the overall accurate glue application of the evaporator.

[0059] In the embodiment, the connecting line between the glue applying point PT1 and the glue applying point PT2 is the glue applying path. It can be understood that the glue applying point PT1 and the glue applying point PT2 are pixel coordinate points, and in order to facilitate the operation of the glue applying robot, the conversion between the pixel coordinate system and the actual coordinate system is realized through the hand-eye calibration in the embodiment.

[0060] In summary, the above only describes the preferred embodiments of the present application, and any equivalent changes and modifications made within the scope of the patent application of the present application shall be covered by the scope of the present application.

Claims

1. A novel method for obtaining the gluing path of an air conditioner evaporator, which is a fusion of two-dimensional and three-dimensional data, characterized by, The method comprises the following steps: Step S1, installing a two-dimensional camera and a three-dimensional camera on an air conditioner evaporator gluing conveying belt, and shooting a brightness image imgA and a depth image imgB of the air conditioner evaporator through the two-dimensional camera and the three-dimensional camera respectively; Step S2, performing binaryzation processing on the depth image imgB to complete segmentation, and obtaining a binaryzation segmented image imgB1; Step S3, performing eight-connected region marking screening on the binaryzation segmented image to obtain a binary image imgB2 after initial screening; Step S4, performing coarse positioning on general plastic parts contained in the screened binary image imgB2 to obtain an image imgB3 after coarse positioning, comprising: Step S41, performing a morphological erosion operation on the screened binary image imgB2 to obtain an eroded binary image; Step S42, performing a morphological dilation operation on the eroded binary image to obtain a dilated binary image; Step S43, performing a connected region screening height operation to screen out a connected region with a height closest to a fixed height of the evaporator plastic part, and obtaining the image imgB3 after coarse positioning; Step S5, performing Gamma correction on the brightness image imgA in step S1 to obtain a corrected image imgA1; Step S6, performing binaryzation processing on the corrected image imgA1, and then using eight-connected region marking screening to obtain an image imgA2; Step S7, subtracting the image imgB3 after coarse positioning in step S4 from the image imgA2 in step S6 to obtain an image of interest imgC; Step S8, performing connected region marking on the image of interest imgC, counting the geometric center of each connected region, and sorting the geometric centers according to the y coordinates from large to small; Step S9, calculating the Euclidean distance between each connected region in step S8 and the connected region labeled as 1, and setting the result as {d1, d2,..., d n} ; taking the connected region with the minimum distance to the preset value T as the target connected region, and the target connected region is the connected region labeled as 3. Step S10, for the connected region with a label of 1, using a least square fitting method to fit a straight line, recording the intersection point of the straight line and the leftmost side of the binary image imgB2 as P1, and recording the coordinates of a point on the right end of P1 on the straight line and having a distance D from P1 as P2; Step S11, for the connected region with a label of 3, using a least square fitting method to fit a straight line, recording the intersection point of the straight line and the leftmost side of the binary image imgB2 as Q1, and recording the coordinates of a point on the right end of Q1 on the straight line and having a distance D from Q1 as Q2; Step S12, obtaining the coordinates of the gluing point: PT1 = w * P1 + (1-w) * Q1, PT2 = w * P2 + (1-w) * Q2, and then obtaining the gluing path of the air conditioner evaporator.

2. A novel method for obtaining the gluing path of an air conditioner evaporator according to claim 1, characterized in that: The two-dimensional camera and the three-dimensional camera in step S1 both adopt an external triggering mode, and the air conditioner evaporator triggers photographing when passing through a flow line photographing station.

3. A novel method for obtaining the gluing path of an air conditioner evaporator according to claim 1, characterized in that: The binaryzation processing in steps S2 and S6 both adopts a fixed threshold value to perform binaryzation segmentation on the depth image, the fixed threshold value is set to 128, and when binaryzation, the pixel values of all pixel points in the depth image imgB are compared one by one, the pixel value of a pixel point is greater than the fixed threshold value, then the pixel point is assigned a value of 1, and the pixel value of a pixel point is less than the fixed threshold value, then the pixel point is assigned a value of 0.

4. The novel method for obtaining the gluing path of the evaporator of an air conditioner based on the fusion of two-dimensional and three-dimensional data according to claim 1, characterized in that: In step S43, the connected region of 80 pixels is selected as the region where the plastic part of the air conditioner evaporator is located.

Images