An Automatic Positioning and Automatic Calibration Method for Elevator Fault Code Identification

Abstract

The present invention discloses an automatic positioning and automatic calibration method for elevator fault code identification, comprising the following steps: A rectangular identification frame with a specific color is arranged around the fault code display screen, and an ambient supplementary light is installed in the elevator control cabinet; The automatic positioning and automatic calibration process is executed. In the first step, RGB image acquisition is performed on the area in the elevator control cabinet that includes the identification frame and the elevator fault code, and one frame of the image is selected; In the second step, automatic positioning processing is performed on the image selected in the first step; In the third step, automatic calibration is performed on the front-facing rectangular RGB image of the fault code display screen. After the equipment is installed using this method, there is no need for manual marking of the position of the elevator fault code display screen, reducing the installation and debugging difficulty.

Description

Technical Field

[0001] The present invention relates to an automatic positioning and automatic calibration method, and particularly to an automatic positioning and automatic calibration method for elevator fault code recognition. Background Art

[0002] At present, for most elevators on the market, the elevator operating status, including door opening / closing status, floor, fault information, etc., is displayed through the elevator fault code display screen on the elevator controller. Generally, the elevator fault code recognition system collects the image of the elevator fault code display screen through an image acquisition module, identifies the content displayed on the fault code display screen in the image through an image processing module, and reports it to the elevator operation and maintenance company and personnel in real time and remotely.

[0003] When realizing the recognition of the content displayed on the fault code display screen, it is first necessary to locate the position area of the fault code display screen and the position of each digit tube in the display screen. Currently, the common method is to manually mark the recognition position area of the fault code display screen in the picture collected by the image acquisition module, and then identify the content displayed on the fault code display screen through an identification algorithm.

[0004] This method of manually calibrating the recognition position area of the fault code display screen has the following problems:

[0005] 1. The on-site implementation process is complex and the workload is large.

[0006] First, the image acquisition module needs to be accurately installed and fixed, the on-site picture needs to be collected, and then uploaded to the computer;

[0007] Secondly, the position of the fault code display screen is manually marked on the computer of the debugging personnel, and then the marked position is configured into the image processing module;

[0008] Finally, it is verified whether the image processing module can correctly identify the fault code displayed on the fault code display screen. If the identification is incorrect, the above process needs to be repeated;

[0009] 2. If problems such as the loosening / reinstallation of the image acquisition module occur, it will cause the previously marked position to not match the position of the fault code display screen in the actual image, exceeding the manually marked frame area, resulting in incorrect identification and the need for re-marking;

[0010] 3. For the image acquisition module installed on the elevator control cabinet door panel, if the door panel is not closed properly, it will cause the angle between the image acquisition module and the elevator fault code display screen to be recognized to change, resulting in the previously marked position not matching the position of the fault code display screen in the actual image, generating deformation, and the fault code to be recognized will leak out of the recognition area, resulting in incorrect identification and the need for re-marking;

[0011] 4. Replacing the image processing module will also cause the problem of marking invalidation;

[0012] 5. When the position of the image acquisition module changes, due to the shooting angle problem, the captured image will be severely deformed. When using the most common wire threading algorithm for recognition, the wire threading recognition area disappears due to the deformation, resulting in inability to recognize and the need for re - marking. Summary of the Invention

[0013] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a method for automatically positioning the fault code display screen and automatically calibrating the recognition area, which can reduce 90% of the preparation workload before fault recognition and can solve the problems of recognition errors and recognition failures caused by manual marking.

[0014] An automatic positioning and automatic calibration method for elevator fault code recognition according to the present invention includes the following steps:

[0015] Step 1: Set a rectangular identification frame with a specific color around the fault code display screen and install an ambient fill light in the elevator control cabinet. The ambient fill light serves as a fill light source for providing the color characteristics of the identification frame.

[0016] If the digital tubes in the fault code display screen are evenly placed in the display area, set an identification frame on the outermost frame of the fault code display screen; if the digital tubes in the fault code display screen are unevenly placed in the display area, first set several adjacent digital tubes with a distance between adjacent digital tubes not exceeding 5 mm as a group of digital tubes and set an identification frame on the outermost frame of each group of digital tubes, and then set an identification frame on the outermost frame of each remaining digital tube.

[0017] Step 2: Execute the automatic positioning and automatic calibration process, which specifically includes the following steps:

[0018] The first step is to perform RGB image acquisition on the area in the elevator control cabinet containing the identification frame and the elevator fault code, and select one frame of the image.

[0019] The second step is to perform automatic positioning processing on the image selected in the first step. The process is as follows:

[0020] Step 201: Use the cvtColor() function of OPENCV to convert the RGB image acquired in the first step into an HSV image, then perform binaryzation on the RGB image according to the HSV image threshold, and then perform erosion and dilation operations on the binaryzation image to remove noise points and connect breakpoints. Use the boundary points for fitting to obtain the quadrilateral area of the identification frame in the entire image. Finally, intercept the fault code display screen image from the RGB image selected in the first step according to the obtained quadrilateral area.

[0021] Step 202: Use the affine transformation algorithm to convert the intercepted image of the fault code display screen into a rectangle, obtaining the front-facing rectangular RGB image of the fault code display screen;

[0022] The third step is to automatically calibrate the front-facing rectangular RGB image of the fault code display screen. The process is as follows:

[0023] For the front-facing rectangular RGB image, calculate the width-to-height ratio value of the image. The rounded integer value obtained by dividing the width-to-height ratio value by 2 / 3 is the number of digital tubes in the image. Divide the rectangular RGB image evenly in the width direction according to the number of digital tubes, and each obtained recognition area image contains only one digital tube.

[0024] The beneficial effects of the present invention are:

[0025] 1. For the work of calibrating the position of the elevator fault code display screen, a large amount of labor costs need to be incurred. Moreover, in order to correctly calibrate, it is also necessary to go to the site to adjust the installation position and angle of the image acquisition module, resulting in low efficiency and a great increase in costs. By using the method of the present invention, about 90% of the relevant expenditure costs can be reduced.

[0026] 2. After the equipment is installed, there is no need to manually mark the position of the elevator fault code display screen, which reduces the installation and debugging difficulty. There is no need for professional personnel to conduct debugging on-site, and the installation and debugging efficiency is reduced from several hours of professional debugging to a few minutes;

[0027] 3. Ensure that the recognition is not affected by the position or shooting angle change of the image acquisition module, and reduce most misrecognition and recognition failure problems.

[0028] 4. After replacing or maintaining the image acquisition module / image processing module, it can be immediately put into use without re-calibration, and the maintenance cost is close to 0;

[0029] 5. Reduce the problems of reduced recognition rate or unrecognizability caused by the inclination and rotation of the image acquisition module and the fault code display screen, resulting in image deformation of the display screen. Description of the Drawings

[0030] Figure 1 is a schematic diagram of the setting of the identification object;

[0031] Figure 2 is a schematic diagram of the setting of the grouped digital tube identification objects;

[0032] Figure 3 is a flowchart of the automatic positioning process;

[0033] Figure 4 is a diagram of the automatic calibration processing result. Detailed Implementation Modes

[0034] To more clearly introduce the objectives, technical solutions, and advantages of the present invention, the present invention will be further described below in conjunction with the accompanying drawings and embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other implementation manners obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the scope protected by the present invention.

[0035] An automatic positioning and automatic calibration method for elevator fault code identification as shown in the accompanying drawings includes the following steps:

[0036] Step 1: Set an identification frame 1 with a specific color and a rectangular shape around the fault code display screen, and install an ambient fill light in the elevator control cabinet. The ambient fill light serves as a fill light source for providing the color characteristics of the identification frame 1.

[0037] The identification frame 1 can be cut from colored raw materials or a specific color can be coated around the identification area using a painting tool. As Figure 1 shown, if the digital tubes in the fault code display screen are evenly placed in the display area, that is, the distance between two adjacent digital tubes is equal and arranged together, then the identification frame 1 is set on the outermost frame of the fault code display screen; if the digital tubes in the fault code display screen are unevenly placed in the display area, then first set several adjacent digital tubes with a distance between adjacent digital tubes not exceeding 5 mm as a group of digital tubes and set an identification frame 1 on the outermost frame of each group of digital tubes, and then set an identification frame 1 on the outermost frame of each remaining digital tube, as shown in the attached Figure 2 figure.

[0038] Step 2: Execute the automatic positioning and automatic calibration process, which specifically includes the following steps:

[0039] The first step is to collect RGB images of the area in the elevator control cabinet containing the identification frame and elevator fault codes, and select one frame of the images.

[0040] The second step is to perform automatic positioning processing on the image selected in the first step, and the process is as follows:

[0041] Step 201: Use the cvtColor() function of OPENCV to convert the RGB image collected in the first step into an HSV image, then perform binaryzation on the RGB image according to the HSV image threshold, and then perform erosion and dilation operations on the binary image to remove noise points, connect breakpoints, use boundary points for fitting to obtain the quadrilateral area of the identification frame in the entire image, and finally intercept the fault code display screen image from the RGB image selected in the first step according to the obtained quadrilateral area.

[0042] In this step, according to the specific color of the identification box 1, the HSV range value of the target color is set, and then the inRange() function of OPENCV is called to extract the color space. This function will convert the remaining colors except the target color into a black background, and only keep this color as the foreground to obtain the extracted HSV image. Then, binary processing is performed on the extracted HSV image, and the gray values of the pixel points on the image are set to 0 or 255, and finally the image only shows obvious black and white. Then, erosion and dilation operations are performed on the binary-processed image to remove noise and connect breakpoints. Finally, the quadrilateral area calibrated by the identification box is obtained by fitting boundary points.

[0043] Through the image processing method in this step, the quadrilateral area of the identification object in the entire image is finally accurately located and recognized, and the RGB image of the fault code display screen is obtained by cropping according to the above quadrilateral area in the original RGB image.

[0044] Step 202, use the affine transformation algorithm to convert the intercepted RGB image of the fault code display screen into a rectangle to obtain the front-view rectangular RGB image of the fault code display screen;

[0045] The third step is to automatically calibrate the front-view rectangular image of the fault code display screen. The process is as follows:

[0046] For the front-view rectangular RGB image, calculate the width-to-height ratio value of the image. The rounded integer value obtained by dividing the width-to-height ratio value by 2 / 3 is the number of digital tubes in the image. The rectangular RGB image is evenly divided in the width direction according to the number of digital tubes, and each obtained recognition area image only contains one digital tube.

[0047] This step is used to determine the number of digital tubes contained in the fault code display screen and the specific positions of each digital tube. The result is to obtain the RGB image of each digital tube.

[0048] The following is an example:

[0049] According to the length-to-width ratio of the corrected rectangular image and referring to the width-to-height ratio of a single digital tube (2:3), through calculation, if the corrected rectangular width-to-height ratio is approximately 2:3, then there is 1 digital tube in this rectangle; if the corrected rectangular width-to-height ratio is approximately 4:3, then there are 2 digital tubes in this rectangle, and so on. For a rectangle containing more than one digital tube, the average method is used to divide it into multiple rectangles, and finally each calibrated recognition area only contains one digital tube. See Figure 4 The calibration processing result.

[0050] Through the above steps, the images of each digit tube in the calibrated fault code display screen can be obtained. Then, as long as a conventional recognition algorithm such as the threading method is used to recognize the images of each digit tube, the characters of each digit of the fault code can be recognized very accurately.

Claims

1. An automatic positioning and automatic calibration method for elevator fault code identification, characterized in that, It is applicable to the working conditions where the inclination and rotation of the image acquisition module and the fault code display screen cause image deformation of the fault code display screen, including the following steps: Step 1: Set an identification frame with a specific color and a rectangular shape around the fault code display screen, and install an ambient fill light in the elevator control cabinet. The ambient fill light serves as a fill light source for providing the color characteristics for identifying the identification frame; the image acquisition module is used to acquire the image of the fault code display screen; If the digital tubes in the fault code display screen are evenly placed in the display area, set an identification frame on the outermost frame of the fault code display screen; if the digital tubes in the fault code display screen are unevenly placed in the display area, first set several adjacent digital tubes with an adjacent digital tube spacing not exceeding 5 mm as a group of digital tubes and set an identification frame on the outermost frame of each group of digital tubes, and then set an identification frame on the outermost frame of each remaining digital tube; Step 2: Execute the automatic positioning and automatic calibration process, which specifically includes the following steps: The first step is to perform RGB image acquisition on the area in the elevator control cabinet that includes the identification frame and the elevator fault code, and select one frame of the image; The second step is to perform automatic positioning processing on the image selected in the first step, and the process is as follows: Step 201: Use the cvtColor() function of OPENCV to convert the RGB image acquired in the first step into an HSV image, then perform binaryzation on the RGB image according to the HSV image threshold, and then perform erosion and dilation operations on the binaryzation image to remove noise points, connect breakpoints, use boundary points for fitting to obtain the quadrilateral area of the identification frame in the entire image, and finally intercept the fault code display screen image in the RGB image selected in the first step according to the obtained quadrilateral area; The specific process of performing binaryzation on the RGB image according to the HSV image threshold in this step is as follows: According to the specific color of the identification frame, set the HSV range value of the target color, and then call the inRange() function of OPENCV to implement the extraction of the color space. This function will convert the remaining colors except the target color into a black background, and only retain this color as the foreground to obtain the extracted HSV image; then perform binaryzation processing on the extracted HSV image, set the grayscale value of the pixel points on the image to 0 or 255, and finally the image presents obviously only black and white; Step 202: Use the affine transformation algorithm to convert the intercepted fault code display screen image into a rectangle to obtain the front-view rectangular RGB image of the fault code display screen; The third step is to perform automatic calibration on the front-view rectangular RGB image of the fault code display screen, and the process is: For the front-view rectangular RGB image, calculate the width-to-height ratio value of the image. The rounded integer value obtained by dividing the width-to-height ratio value by 2 / 3 is the number of digital tubes in the image. Divide the rectangular RGB image evenly in the width direction according to the number of digital tubes, and each obtained recognition area image only contains one digital tube.

Images