Automatic testing system and method for automobile instrument indicator lamp

Abstract

The invention provides an automatic testing system and method for an automobile instrument indicating lamp, and belongs to the technical field of automobile display screen testing. The automatic testing system comprises an automobile instrument, a signal simulation module, an image acquisition module and a testing module, and the output end of the signal simulation module is connected with the input end of the automobile instrument; the output end of the signal simulation module is connected with the input end of the test module, the image acquisition module is used for acquiring images of the automobile instrument, and the output end of the image acquisition module is connected with the input end of the test module. During working, the signal simulation module inputs a simulation signal to the automobile instrument, and the automobile instrument displays an image according to the simulation signal. The image acquisition module acquires an image displayed by the automobile instrument and transmits the image to the test module. And the test module performs model reasoning on the image through a target detection algorithm, and compares a model reasoning result with an expected result to complete assertion. The accuracy of the test result is improved, the test cost is reduced, and the test efficiency is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of automobile display screen testing, and in particular, relates to an automatic testing system and method for automobile instrument indicator lights. Background Art

[0002] As cars develop towards intelligent networking, smart cockpit systems are becoming more and more popular in passenger cars. Smart cockpit systems consist of two subsystems: the car central control and the car instrument. The car instrument can display the speedometer, tachometer, oil pressure gauge, water temperature gauge, fuel gauge, and various indicator lights (such as turn signals, seat belt unfastened indicator lights, tire pressure fault lights, water temperature working warning lights, etc.), providing drivers with the required car operating parameter information. There are many types of car instrument indicator lights, which account for a large proportion of car instrument testing.

[0003] Chinese Patent 101488028A provides a system and method for testing CAN bus automobile instruments, the system includes a configuration unit, an input unit, a virtual output unit and a comparison unit; the method includes the following steps: Step 31, in the configuration unit, configure the control logic of the CAN bus automobile instrument; Step 32, the input unit sends a CAN message to the CAN bus automobile instrument and the virtual output unit; Step 33, the virtual output unit simulates the output of the CAN bus automobile instrument according to the control logic in the configuration unit and the CAN message sent by the input unit; Step 34, the comparison unit detects whether the output of the virtual output unit is consistent with the output of the CAN bus automobile instrument; if they are consistent, go to step 35; if they are inconsistent, go to step 36; Step 35, the CAN bus automobile instrument test is correct; Step 36, the CAN bus automobile instrument test is wrong; Step 37, the test process ends.

[0004] The existing technology has a complex testing process for automobile instruments and low testing efficiency. Summary of the invention

[0005] The present invention aims to provide an automatic testing system and method for automobile instrument indicator lights, so as to achieve the technical purpose of improving the testing efficiency of automobile instruments.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is:

[0007] The present invention provides an automatic test system for an automobile instrument indicator light, comprising an automobile instrument, a signal simulation module, an image acquisition module and a test module, wherein the output end of the signal simulation module is connected to the input end of the automobile instrument, the output end of the signal simulation module is connected to the input end of the test module, the image acquisition module is used to acquire an image of the automobile instrument, and the output end of the image acquisition module is connected to the input end of the test module.

[0008] The signal simulation module inputs a simulation signal to the automobile instrument, the image acquisition module inputs the acquired image information to the test module, and the signal simulation module outputs the simulated test information to the test module.

[0009] The signal simulation module adopts a CAN bus simulator.

[0010] The image acquisition module adopts a high-definition camera.

[0011] The test module adopts a computer.

[0012] The present invention provides an automatic testing method for an automobile instrument indicator light:

[0013] Step 1: The signal simulation module inputs the analog signal to the car instrument, and the car instrument displays the image according to the analog signal;

[0014] Step 2: The image acquisition module acquires the image displayed by the vehicle instrument and transmits the image to the test module;

[0015] Step 3: The test module performs model inference on the image through the target detection algorithm, and compares the result of model inference with the expected result to complete the assertion.

[0016] In step 1, the signal simulation module simulates the CAN signal and inputs it into the vehicle instrument.

[0017] In step 2, the high-definition camera transmits the image information of the vehicle instrument to the computer connected to the high-definition camera by shooting.

[0018] In step three, the test module obtains training images through a high-definition camera before applying the target detection algorithm, annotates the data according to the instrument category and indicator light category, then builds a model training environment, starts model training, verifies the model reasoning effect, and continuously iterates the model by supplementing training images and modifying data annotations, finally obtaining a model for testing.

[0019] The technical effects of the present invention are:

[0020] (1) The present invention uses a target detection algorithm to test automobile instruments, which reduces the requirements for the quality of images collected by the camera, thereby reducing the impact of the environment on the test results, increasing the accuracy of the test results, and reducing the test cost.

[0021] (2) The present invention collects information of automobile instruments through a high-definition camera, and the information collection is simpler and more direct. It is also suitable for automobile instruments without a screenshot function and has a wider range of applications.

[0022] (3) The target detection algorithm used in the present invention has a fast detection speed and high detection efficiency, which meets the test speed requirements of large-scale automobile instrument testing.

[0023] (4) The present invention improves the model reasoning capability and increases the accuracy of the test by continuously training the target detection algorithm model. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] This specification includes the following drawings, which show the following contents:

[0025] Figure 1 A logical structure diagram of an automatic test system and method for an automobile instrument indicator light of the present invention;

[0026] Figure 1 The markings are: 1. Automobile instrument; 2. Signal simulation module; 3. Image acquisition module; 4. Test module. DETAILED DESCRIPTION

[0027] The specific implementation methods of the present invention are further explained in detail below by describing the embodiments with reference to the accompanying drawings, with the aim of helping those skilled in the art to have a more complete, accurate and in-depth understanding of the inventive concept and technical solution of the present invention and facilitating their implementation.

[0028] The present invention provides an automatic test system for an automobile instrument indicator light, comprising an automobile instrument 1, a signal simulation module 2, an image acquisition module 3 and a test module 4, wherein the output end of the signal simulation module 2 is connected to the input end of the automobile instrument 1, the output end of the signal simulation module 2 is connected to the input end of the test module 4, the image acquisition module 3 is used to acquire an image of the automobile instrument 1, and the output end of the image acquisition module 3 is connected to the input end of the test module 4. The signal simulation module 2 inputs an analog signal to the automobile instrument 1, the image acquisition module 3 inputs the acquired image information to the test module 4, and the signal simulation module 2 outputs the information of the analog test to the test module 4.

[0029] The signal simulation module 2 uses a CAN bus simulator. The image acquisition module 3 uses a high-definition camera. The test module 4 uses a computer.

[0030] The present invention provides an automatic testing method for an automobile instrument indicator light:

[0031] Step 1: The signal simulation module 2 inputs the analog signal to the automobile meter 1, and the automobile meter 1 displays an image according to the analog signal.

[0032] Step 2: The image acquisition module 3 acquires the image displayed by the automobile instrument 1 and transmits the image to the testing module 4 .

[0033] Step 3: Test module 4 performs model inference on the image through the target detection algorithm, and compares the result of the model inference with the expected result to complete the assertion.

[0034] In step 1, the signal simulation module 2 simulates the CAN signal and inputs it into the automobile instrument 1.

[0035] In step 2, the high-definition camera transmits the image information of the automobile instrument 1 to the computer connected to the high-definition camera by shooting.

[0036] In step three, the test module 4 obtains the training images through the high-definition camera before applying the target detection algorithm, annotates the data according to the instrument category and the indicator light category, then builds the model training environment, starts model training, verifies the model reasoning effect, and continuously iterates the model by supplementing training images and modifying data annotations, and finally obtains the model for testing.

[0037] The following is a detailed description of an automatic testing system and method for an automobile instrument indicator light of the present invention.

[0038] The present invention discloses an automatic test system for an automobile instrument indicator light, comprising an automobile instrument 1, a signal simulation module 2, an image acquisition module 3 and a test module 4, wherein the automobile instrument 1 is a device to be tested, and the automobile instrument 1 displays image information of a speedometer, a tachometer, an oil pressure gauge, a water temperature gauge and a fuel gauge, as well as a turn signal, a seat belt unfastened indicator light, a tire pressure fault light and a water temperature working warning light. A user can understand the state of the automobile through this information, and the automobile instrument 1 is very important for the driving safety of the user.

[0039] The signal simulation module 2 uses a CAN bus simulator to send a simulation signal to the automobile meter 1, so as to simulate the signal collected by the automobile sensor and input it into the automobile meter 1. The instrument and indicator light on the automobile meter 1 display the information collected by the automobile sensor. In the present invention, the automobile meter 1 displays the information collected by the automobile sensor simulated by the signal simulation module 2. In the embodiment of the present invention, the CAN bus simulator is ICSim.

[0040] The image acquisition module 3 uses a high-definition camera to capture the automobile instrument 1 to obtain an image, and input the image into the test module 4. The present invention collects information of the automobile instrument 1 through a high-definition camera, and the information collection is simpler and more direct. It is also applicable to automobile instruments 1 without a screenshot function, and has a wider range of applications.

[0041] The test module 4 uses a computer, and the test module 4 is used to test the function of the automobile instrument 1 and train the algorithm model before the test. Specifically, before the test, the test module 4 collects the image data of the automobile instrument 1 through the image acquisition module 3 as the training image, and the test module 4 marks the instrument category and indicator light category in the image, determines the category and boundary box of the instrument and indicator light, and then selects the target detection algorithm to build the model training environment, wherein the target detection algorithm can select YOLO, SSD, or Faster R-CNN algorithm. The test module 4 trains the model and verifies the reasoning effect of the model, that is, the reasoning result of the model is compared with the prediction result input by the signal simulation module 2 to determine the effect of the model reasoning, and then the model is continuously iterated by supplementing the training image and modifying the data annotation, so as to obtain the final model for testing. After obtaining the model for testing, the test module 4 starts to test the automobile instrument 1. Specifically, the test module 4 collects the image of the automobile instrument 1 through the image acquisition module 3, and infers the image through the target detection algorithm to obtain the reasoning result, and compares the reasoning result with the prediction result input by the signal simulation module 2 to complete the assertion. The target detection algorithm adopted in the present invention has fast detection speed and high detection efficiency, and meets the test speed requirements of large-scale automobile instrument 1 testing.

[0042] The output end of the CAN bus simulator is connected to the input end of the automobile instrument 1, the output end of the CAN bus simulator is connected to the input end of the computer, and the output end of the high-definition camera is connected to the input end of the computer.

[0043] The present invention provides an automatic testing method for an automobile instrument indicator light:

[0044] Before testing, the model of the target detection algorithm in the present invention needs to be trained. Specifically, the test module 4 collects the image of the automobile instrument 1 through the image acquisition module 3 and inputs it to the test module 4. The test module 4 marks the instrument category and the indicator light category in the image, and then builds a model training environment. The test module 4 trains the model to verify the reasoning effect of the model, compares the reasoning result of the model with the prediction result input by the signal simulation module 2 to determine the effect of the model reasoning, and then continuously iterates the model by supplementing the training image and modifying the data annotation to obtain a model for testing. The present invention improves the ability of model reasoning and increases the accuracy of the test by continuously training the target detection algorithm model.

[0045] After obtaining the model of the target detection algorithm for testing, the system starts testing. The signal simulation module 2 inputs the simulation signal to the automobile instrument 1, and the automobile instrument 1 displays the image according to the simulation signal. To illustrate with an embodiment, the signal simulation module 2 inputs the automobile left turn signal on signal and the seat belt unfastened signal to the automobile instrument 1, and the left turn signal on the automobile instrument 1 lights up, and the seat belt unfastened indicator light lights up.

[0046] The image acquisition module 3 acquires the image information of the automobile instrument 1 and inputs it to the test module 4. The target detection algorithm of the test module 4 performs model inference on the image, and compares the result of the model inference with the expected result to complete the assertion. In an embodiment of the present invention, the test module 4 infers the category and coordinate information of the instrument and indicator light, and compares the inferred category and coordinate information of the instrument and indicator light with the simulated test information input by the signal simulation module 2, that is, the test module 4 infers the lighting information and coordinate information of the left turn signal and the seat belt unfastened indicator light, and compares it with the predicted result input by the signal simulation module 2 to complete the assertion. The assertion is to verify the expected result. If the result inferred by the target detection algorithm is the lighting information and coordinate information of the left turn signal and the seat belt unfastened indicator light input by the signal simulation module 2, then the corresponding automobile instrument 1 is detected. The present invention is different from the traditional detection method that only relies on the camera to take pictures and then perform manual comparison. The traditional method is heavily dependent on the distance, angle and ambient light between the camera and the instrument when shooting, and has low anti-interference ability to the environment. The present invention can cope with tests in different environments through a large amount of data training model.

[0047] The present invention tests the automobile instrument 1 through a target detection algorithm, which reduces the requirements on the quality of images collected by the camera, thereby reducing the impact of the environment on the test results, increasing the accuracy of the test results, and reducing the test cost.

[0048] The present invention is described above by way of example in conjunction with the accompanying drawings. Obviously, the specific implementation of the present invention is not limited to the above-mentioned methods. As long as various non-substantial improvements are made using the method concept and technical solution of the present invention; or the above concept and technical solution of the present invention are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.

Claims

1. An automatic test system for automobile instrument indicator lights, characterized in that: It includes an automobile instrument, a signal simulation module, an image acquisition module and a test module. The output end of the signal simulation module is connected to the input end of the automobile instrument, the output end of the signal simulation module is connected to the input end of the test module, the image acquisition module is used to acquire images of the automobile instrument, and the output end of the image acquisition module is connected to the input end of the test module.

2. The automatic test system for automobile instrument indicator lights according to claim 1, characterized in that: The signal simulation module inputs a simulation signal to the automobile instrument, the image acquisition module inputs the acquired image information to the test module, and the signal simulation module outputs the simulated test information to the test module.

3. The automatic test system for automobile instrument indicator lights according to claim 1, characterized in that: The signal simulation module adopts a CAN bus simulator.

4. The automatic test system for automobile instrument indicator lights according to claim 1, characterized in that: The image acquisition module adopts a high-definition camera.

5. The automatic test system for automobile instrument indicator lights according to claim 1, characterized in that: The test module adopts a computer.

6. An automated testing method for an automobile instrument indicator light according to any one of claims 1 to 5, characterized in that: Step 1: The signal simulation module inputs the analog signal to the car instrument, and the car instrument displays the image according to the analog signal; Step 2: The image acquisition module acquires the image displayed by the vehicle instrument and transmits the image to the test module; Step 3: The test module performs model inference on the image through the target detection algorithm, and compares the result of model inference with the expected result to complete the assertion.

7. The method for automatically testing an automobile instrument indicator light according to claim 6, characterized in that: In step 1, the signal simulation module simulates the CAN signal and inputs it into the vehicle instrument.

8. The method for automatically testing an automobile instrument indicator light according to claim 6, characterized in that: In step 2, the high-definition camera transmits the image information of the vehicle instrument to the computer connected to the high-definition camera by shooting.

9. The method for automatically testing an automobile instrument indicator light according to claim 6, characterized in that: In step three, the test module obtains training images through a high-definition camera before applying the target detection algorithm, annotates the data according to the instrument category and indicator light category, then builds a model training environment, starts model training, verifies the model reasoning effect, and continuously iterates the model by supplementing training images and modifying data annotations, finally obtaining a model for testing.

Images