A FAULT DETECTION SYSTEM

Abstract

This invention relates to a system (1) that enables real-time damage detection on the outer surfaces of vehicles passing through a tunnel containing camera (4) and lighting elements (3) using artificial intelligence algorithms.

Description

1 TARIFF A FAULT DETECTION SYSTEM Technical Area 5 This invention allows vehicles to be driven through a tunnel containing camera and lighting elements. Real-time defect detection on their outer surfaces using artificial intelligence algorithms. It is related to a system that enables this to happen. Previous Technique Damage assessment on the exterior of vehicles, whether new or used, is done manually. This is done through examination or traditional image processing techniques. Today With existing methods, damage assessment is both time-consuming and prone to human error. 15 This situation involves incorrect costing and errors in maintenance and insurance processes. This leads to reporting. On the other hand, damage assessment usually requires fixed points. Images captured from camera angles or mobile devices can be easily analyzed. Simple image processing or computer vision methods are used. The solutions used enable high-resolution inspection of vehicle surfaces. 20 It does not present any damage. Furthermore, damage is repeatedly visible in images taken from different angles. This can lead to duplicate error detections and counts. Therefore, today it will offer the possibility of high-resolution examination, in different frames. a set of accurate, fast, and standardized decisions that can uniquely identify the same damage. Solutions are needed to provide damage assessment. 25 European patent number EP3696537B1, which is included in the prior art. The document mentions a damage assessment device and system for vehicles. The invention describes a device for detecting damage to a moving vehicle. is provided. The device provides lighting for illuminating a moving vehicle. 30 The system uses a sensor system to receive image data from the vehicle and the received image. 2 An assessment system that determines vehicle damage based on data. It includes. The lighting system includes at least one of the following elements: for illuminating a moving vehicle with monochromatic and / or polarized light a lighting unit configured as such and illuminating the vehicle with the primary light A primary lighting unit configured for this purpose. The sensor system is 5. It includes at least one of the following elements: from a moving vehicle, Acquiring image data when illuminated with monochromatic and / or polarized light a sensor element configured for this purpose; also with the first light When illuminated, the first moving vehicle in a specific field of view image data and the same image when illuminated with a second light source 10 one or more configured to acquire secondary image data in the field. Multiple sensor elements. Here, the second lighting is created by the first light. It is different from lighting. Brief Description of the Invention 15 The purpose of this invention is to provide applications in the automotive, logistics, insurance, leasing, and mobility industries. visual data collected from vehicles is processed using artificial intelligence algorithms. the examination and analysis of damage to the exterior surfaces of the vehicles The goal is to create a system that enables the identification and classification of these data. 20 The other purpose of the invention is to protect vehicles, especially those to be delivered as brand new, before delivery. by being subjected to pre-delivery inspection (PDI) on their outer surfaces The goal is to create a system that enables real-time damage assessment. Another purpose of the invention is to improve quality control, which is manual and prone to error. making processes more efficient with an AI-based solution, saving time and effort The goal is to create a system that provides benefits in terms of power. 3 Another purpose of the invention is to enable the automation of quality control processes. increasing efficiency and transferring the identified errors to the production process. The goal is to create a system that enables the improvement of these stages. Another purpose of the invention is to record images of vehicles that are ready for delivery. the automatic receipt of the pre-delivery inspection checklist a system that enables filling and integration into internal system platforms to accomplish. Another purpose of the invention is to automatically assess the damage if damage is detected. The goal is to create a system that enables the creation of work orders to resolve the issue. Detailed Description of the Invention To achieve the purpose of this invention, a "Fault Detection System" was developed (Appendix 15). as shown in the figures, these figures are: Figure 1. Appearance of a fault detection system, which is the subject of the invention. Figure 2 shows the appearance of the AA detail in Figure 1. The parts shown in the figures are individually numbered, and these numbers... The equivalents are given below: 1. System 2. Tunnel 25 3. Lighting element 4. Camera 5. Server 30 visuals collected from vehicles, used in the automotive and mobility industries. The external analysis of the data using artificial intelligence algorithms revealed the vehicles' external capabilities. 4 detection, analysis and classification of defects on their surfaces a system that provides the subject of the invention (1), -the vehicle being tested for errors passes through it, and the errors on the vehicles are compared with contrast. The best one with a zebra-patterned surface on the inside to make it more distinctive. a small tunnel (2), 5 - located on the vehicle-facing surface of the tunnel (2) and at least on the vehicle surface at least one configured to provide lighting to highlight errors lighting element (3), -the zebra pattern surface on the inner surface of the tunnel (2) facing the vehicles and The lighting element (3) is 10 meters high from the vehicle passing through the tunnel (2) at least one camera (4) configured to collect images at resolution and -to communicate with the camera (4) through any communication protocol, established to exchange data via communication, collected by camera (4) processing images, running deep learning models, and on the vehicle surface The errors are identified, analyzed, and analyzed using location, size, and time information. 15 To ensure classification, the same error can be detected differently from multiple cameras (4) If the times are detected, the same errors will occur with location, size and time information. by using and unifying them under a single identity and through an interface and configured to generate a report on defects on the vehicle surface It contains a small number of servers (5). 20 The tunnel (2) in the system in question (1) is in the shape of an inverted U and from its ends It is fixed to the area where it will be used and has lighting elements on the surface facing the ground. (3) and includes cameras (4). In the preferred application of the invention, the tunnel (2) The surface facing the ground, i.e., the vehicle, has a zebra-patterned surface. This allows for 25 with the illumination of the lighting elements (3) distortions, damage on the vehicle surface etc. The errors become more apparent with contrast. The stripes are related to zebra patterns. For passive deflectionometry, the tunnel (2) is positioned by imprinting it on the inner surface. The tunnel (2) has a modular structure and a lightweight but durable body structure. It has. It can also be easily mounted in the area where it will be used and is 30 It can be disassembled. The lighting elements in the system (1) that is the subject of the invention are (3) white LEDs. There are projectors that emit light from the inner surface of the tunnel (2) towards the vehicle and the vehicle Consistent and controlled lighting to highlight surface imperfections. It provides (3) temporal modulation in lighting elements. There is no lighting element (3) and the use of zebra pattern makes the vehicle 5 This allows surface damage to be seen more clearly. Additionally, it provides lighting. elements (3) to create homogeneous light with passive printed strips It is being structured. The camera (4) in the system (1) which is the subject of the invention is used for industrial purposes. The image collection devices are located on the tunnel (2) and move continuously. To view the vehicle surface of vehicles from multiple angles and in high-resolution images. It is used to collect images at high resolution. Camera (4) has more than one numerous units collect data by simultaneously and thoroughly scanning all parts of the vehicle. and is configured to share with the server (5). The camera (4) is unfiltered 15 Independent monochrome cameras with fixed lenses are used. Preferably in a tunnel. (2) There are physical slots for up to (4) ten cameras. The cameras (4) are single spectral, It is configured to collect unpolarized images. In the system that is the subject of the invention, the server (5) is a processing unit and the cameras (4) 20 to communicate, to exchange data through the established communication, processing images, running deep learning models, and performing error analysis. It is configured to enable its implementation. The server (5) also has a information about an electronic device and / or the vehicle being fault-detected via the interface On the screen, it presents the user with reports regarding errors on the vehicle surface, analysis 25 It is structured to transmit results and deduplicated data. Server (5) noise reduction and edge enhancement in collected image data performing preliminary processes in this way and using deep learning models to develop artificial intelligence. By performing a damage analysis based on 30 types of damage, the vehicle can be identified as scratches, dents, and paint defects. to detect errors and label them with information on location, size, color, and time. 6 is being configured. Server (5) the same error occurs on multiple cameras (4) and / or Location, size, color, and time when detected at different time frames using their information to unify the same damage and combine it into a single identity. It is structured in such a way that it enables the detection of repetitive errors. This prevents the reporting of false positives and improves the accuracy of reports. Server 5 (5) also regarding the coverage logic, the overlap is offline with homography. This approach eliminates concerns about parallax. Server (5) simultaneous classification with single-stream convolutional neural networks (CNN) and Performing pixel-based segmentation, removing defects to reduce duplicates. clustering spatio-temporal and reporting when the vehicle exits the tunnel (2) It is configured to share. The server (5) preferably has a compressed annotated version. It shares images and JSON data with the user. Server (5) vehicle information display and / or user electronic 15 in preferred application to communicate with the device via any communication protocol and the established After error analysis via communication, the results obtained are presented through an interface. It is configured to share. The server (5) communicates with external servers, It is configured to share error reports and analyses. Server (5) 20 It is structured in such a way that it provides deduplicated error analysis results. accurate costing, maintenance planning, and insurance reporting It is possible. Server (5) processes the collected visual data and conducts a pre-delivery inspection (PDI) 25 The list is automatically generated based on the review items included in the checklist. to ensure it is filled in, automatically share it with external servers and work It is configured to enable the activation of the command. This allows the vehicle to... If an error is detected, a work order is opened to correct the relevant error. The process can be accelerated. 30 7 Server (5) to perform fully automated reporting and preferably to insurance and fleet systems. by integrating it to enable the sharing of error reports It is configured. The server (5) also sends reports to cloud or local servers. It can share. Server (5) segmentation, object detection and hybrid models Detecting errors inside and outside the vehicle, and continuous learning and data updating 5 by performing the work on-site or in the cloud. is configured. Server (5) for in-vehicle fault detection. using photographs and artificial intelligence in elements such as the cabin and trim components to perform checks and track internal hardware version control. It is being structured. 10 Thanks to the system (1) that is the subject of the invention, tunnel (2) can be used in the automotive and mobility industry. Scratches, dents, and paint defects on the exterior surface of vehicles that move continuously. Errors such as these are detected and analyzed using intelligence-based algorithms, and It is classified. Cameras on vehicles passing through tunnel (2) in the system (1) 15 Images are collected via (4). Images that cameras (4) will collect By contrasting with the LED lighting elements (3) and zebra pattern surface, It is being clarified and collected as highly accurate visual data. Server (5) Using artificial intelligence techniques, the collected data can be analyzed to identify scratches specific to each brand of vehicle. It detects and identifies dents and paint defects. The server (5) detected 20 If errors are detected from different angles and in multiple frames, then the errors are identified. It unifies the control processes. This allows for control processes that were previously carried out manually. It is being automated and AI-based solutions are improving efficiency, time, and labor. The benefit is provided. The system (1) is also used in quality control areas. It contributes to the automation of quality processes, increases efficiency, and detects 25 The errors detected are shared with the production system instantly and automatically. On the other hand, the system (1) is used in service areas before delivery of vehicles Mandatory pre-delivery inspection (PDI) checks are AI-based. to ensure that this is done and that images of the vehicles are recorded during the delivery phase With the receipt, the PDI checklist can be filled in automatically. In the system (1) 30 8 By integrating the server (5) into the internal system platforms, there is no damage to the vehicle. If detected, an automatic work order can be created. Around these basic concepts, the subject of the invention is "An Error Detection System (1)". It is possible to develop a wide variety of applications, and the invention is one of the 5 described herein. It cannot be limited to examples; it is essentially as stated in the requests.

Claims

9 REQUESTS 1. Visuals collected from vehicles, used in the automotive and mobility industries. The external analysis of the data using artificial intelligence algorithms revealed the vehicles' external capabilities.

5. Detecting, analyzing, and classifying defects on their surfaces providing, -the vehicle being tested for errors passes through it, and the errors on the vehicles are compared with contrast. The best one with a zebra-patterned surface on the inside to make it more distinctive. a small tunnel (2), - located on the vehicle-facing surface of the tunnel (2) and at least 10 on the vehicle surface at least one configured to provide lighting to highlight errors lighting element (3), -the zebra pattern surface on the inner surface of the tunnel (2) facing the vehicles and Higher than the vehicle passing through the tunnel (2) where the lighting element (3) is used at least one camera (4) configured to collect images at resolution 15 -to communicate with the camera (4) through any communication protocol, established to exchange data via communication, collected by camera (4) processing images, running deep learning models, and on the vehicle surface detecting and analyzing errors with information on location, size, and time, and to ensure classification, the same error from multiple cameras (4) different 20 If the times are detected, the same errors will occur with location, size and time information. by using and unifying them under a single identity and through an interface and configured to generate a report on defects on the vehicle surface a system characterized by a small server (5) (1).

2. It is in the shape of an inverted U, and its ends are fixed to the area where it will be used and placed on the ground. tunnel (2) with lighting elements (3) and cameras (4) on the viewing surface A system like the one in Claim 1, characterized (1).

3. The inner surface of the stripes relating to zebra patterns for passive deflectionometry. Claim 1 is characterized by the tunnel (2) in which it is positioned in a printed manner or A system like the one in 2 (1).

4. White LED projectors illuminate the inner surface of the tunnel (2) towards the vehicle. 5 a consistent and controlled way to highlight and accentuate imperfections on the vehicle surface The above is characterized by lighting elements (3) that provide illumination. a system like any of the requests (1).

5. Configured to create homogeneous light with passively printed strips. any of the above requirements characterized by lighting elements (3) a system like one of them (1).

6. Image acquisition devices used for industrial purposes and tunnels (2) located on, the vehicle surface of continuously moving vehicles multiple 15 used to view from an angle and to collect high-resolution images. as in any of the above requests characterized by camera (4) a system (1).

7. Multiple in number, and all sides of the vehicle simultaneously and in detail 20 camera configured to collect data by scanning and share it with the server (5) A system like any of the above-mentioned demands characterized by (4). (1).

8. It is a processing unit and communicates with cameras (4), the established communication 25 exchanging data, processing images, deep learning to run the models and enable error analysis any of the above requests characterized by the configured server (5) a system like one of them (1). 11 9. An interface is used to detect errors in an electronic device and / or perform fault detection. To present the user with reports on defects on the vehicle's surface on the vehicle's information display. Server configured to transmit analysis results and deduplicated data (5) a system like any of the above-mentioned claims characterized by (1). 5 10. Noise reduction and edge enhancement in collected image data. performing preliminary processes in this way and using deep learning models to develop artificial intelligence. By performing a damage analysis based on data such as scratches, dents and paint defects on the vehicle. 10 to detect errors and label them with information on location, size, color, and time. any of the above requests characterized by the configured server (5) a system like one of them (1).

11. The same error in more than one camera (4) and / or different time frames If detected, the same 15 will be used to track location, size, color, and time information. Server configured to unify the damage and combine it into a single identity (5) a system like any of the above-mentioned claims characterized by (1).

12. Simultaneous classification and pixel 20 with single-stream convolutional neural networks (CNN). Performing segmentation based on defects to reduce duplicates clustering spatio-temporal and reporting when the vehicle exits the tunnel (2) The above is characterized by the server (5) configured to share. a system like any of the requests (1).

13. Any communication with the vehicle information display and / or user electronic device. Communicating via the protocol and performing error analysis based on the established communication. then configured to share the results obtained through an interface as in any of the above requests characterized by the server (5) a system (1). 30 12 14. Communicating with external servers, sharing bug reports and analyses. from the above requests characterized by the server (5) configured for a system like any other (1).

15. Integrated into vehicle production lines and / or maintenance, insurance, and inspection processes. The above is characterized by the server (5) which is configured to be able to do so. a system like any of the requests (1).

16. Process the collected visual data for a pre-delivery inspection (PDI) check. Depending on the review items included in the list, the list is automatically expanded to 10. to ensure completion, automatically share with external servers and the work order characterized by the server (5) configured to enable opening a system like any of the above requests (1).

17. Segmentation, object detection, and hybrid models with 15 features on the interior and exterior of the vehicle. by detecting errors and continuously learning and updating data on-site or Server (5) configured to work in the cloud (onprem / cloud) a system like any of the above characterized demands (1).

18. Using interior photos of the vehicle to diagnose interior faults and the cabin, 20 AI-controlled features such as trim components and interior fittings. Characterized by the server (5) configured to perform version tracking. a system like any of the above-mentioned requests (1).

Images