Recording device

Abstract

The present application relates to a recording device according to the invention, which comprises an analysis device with an analysis location at which a vehicle can be positioned, wherein the analysis device comprises at least one, for example exactly one, sensor directed at the analysis location, wherein the sensor is configured to generate sensor data that characterize one or more states of the vehicle when the vehicle is positioned at the analysis location, wherein the sensor comprises at least one camera, one laser, one mechanical device, one touch sensor, one liquid sensor, one infrared sensor and combinations thereof;and a computing unit that is communicatively connected to the sensor and includes a memory of instructions which, when executed, cause the computing unit to initiate the execution of an inspection algorithm that automatically performs an inspection of the vehicle in response to the vehicle's reception at the analysis location, the computing unit automatically controlling the operation of the sensor during the execution of the inspection algorithm.

Description

The present application relates according to the invention to a recording device comprising an analysis device having an analysis location at which a vehicle can be positioned, the analysis device comprising at least one, for example exactly one, sensor directed towards the analysis location, the sensor being configured to generate sensor data characterizing one or more states of the vehicle when the vehicle is positioned at the analysis location, the sensor comprising at least one camera, a laser, a mechanical device, a tactile sensor, a liquid sensor, an infrared sensor and combinations thereof; and a computing unit communicatively connected to the sensor and comprising a memory with instructions which, when executed, cause the computing unit to initiate execution of an inspection algorithm, which automatically performs an inspection of the vehicle in response to the receipt of the vehicle at the analysis location, wherein the computer unit automatically controls the operation of the sensor during the execution of the inspection algorithm, and the receipt of the sensor data of the sensor, wherein the sensor data characterize the one or more states of the vehicle, and the execution of an evaluation algorithm which causes the computing device to automatically analyze the sensor data and to automatically identify damage to the vehicle on the basis of the sensor data, and the execution of an estimation algorithm which causes the computing device to automatically estimate the damage variable, inter alia therewith, or an area extent on a body surface of the car and / or a depth shift and / or depth change profile of the body surface of the car, for auxiliary units such as a fire weir, THW, The method of performing police and / or the in particular technical scope of the repair of the vehicle, which is based at least in part on the identified damage, wherein the estimation algorithm is based on or is based on a framework for data stream-oriented programming, such that by machine learning, in particular by. Arithmetic operations which are carried out by artificial neural networks on multidimensional data fields, so-called tensors, an artificial intelligence (AI) can be generatedThe present invention seeks to provide a vehicle owner with a service to be able to detect and evaluate damage to his vehicle in real time so that the owner can immediately take an image of the extent and cost of the damage. Auto-rental companies or insurance companies are being offered a AI tool that rationalizes damage processing so that these damage can be regulated more quickly and more accurately.In addition, the user is provided with a guided process for capturing the damage to ensure that the images provided to the user, for example, an insurance company, are accurate and useful for assessment and that the service provides a detailed description of the damage, including various repair methods, to allow the full extent of the damage and the steps for repair to be understood. In addition, the service provides an integral damage calculation that includes spare parts, labor, and paintwork to prepare for repair costs. Auto-rental or insurance companies wish the AI tool to be based on SilverDA or similar computation and current vehicle and repair data in order for the damage computation to be accurate and current.By creating a dataset specifically targeted for medium damage, a wider range of cases can be detected and the AI trained to better detect and classify such cases. This results in higher accuracy and more reliable predictions. In contrast to open source data sets, the present invention has full control over its content and quality when creating its own data sets. Images with different environmental conditions, lighting conditions, perspectives, and damage patterns can be carefully compiled, and it is thus ensured that the AI is trained on the most relevant and representative data.Open source data sets often reflect the real conditions only to a limited extent. Creating the dataset of the present invention allows images from separate projects to be included, thus taking into account the unique challenges and peculiarities of the respective organization. This approach in practice will enable the AI to master situations specific to our operation.If the customer owns the record, he may continually improve him over time. As the AI continues to develop and new types of damage occur, the dataset can be easily updated and augmented so that present AI can adjust and improve its detection capabilities accordingly.According to at least one embodiment, the present invention provides a system wherein the memory further causes the computing device to receive, from a device configured to automatically detect vehicle accidents, a first damage indication related to the vehicle in response to the device detecting an accident involving the vehicle, the first damage indication including telematics information collected by the device during operation of the vehicle.According to at least one embodiment, telematics information includes one or more speed measurements, one or more brake measurements, and one or more acceleration measurements of the vehicle either before, after, or before and after a time of the accident, wherein telematics information includes one or more vehicle parameters collected via an on-board diagnostic system (OBD) connected to the vehicle.According to at least one embodiment, the one or more vehicle parameters include one or more measurements of mileage, coolant temperature, engine speed, spark timing, throttle position, and one or more measurements from an in-vehicle oxygen sensor.According to at least one embodiment, the present invention provides a non-transitory computer readable storage medium having stored thereon computer executable program instructions which, when executed by a processor, cause a computing device to (i) initiate execution of an inspection algorithm that automatically performs inspection of a vehicle in response to receipt of data by the sensors:(i) initiating execution of an inspection algorithm that automatically performs inspection of a vehicle in response to receipt of the vehicle at an analysis location within an analysis device, the analysis device comprising at least one camera, for example, exactly one camera, directed at the analysis location, the camera configured to generate data characterizing one or more conditions of the vehicle when the vehicle is positioned at the analysis location, and wherein the camera is communicatively coupled to the computing device and the computing device automatically controls operation of the camera during execution of the inspection algorithm; (ii) receiving the data characterizing the one or more conditions of the vehicle from the camera; (iii) executing an evaluation algorithm that causes the computing device to automatically analyze the data and to automatically identify damage to the vehicle based on the data; and (iv) executing an estimation algorithm that causes the computing device to automatically make an estimate of the damage magnitude for auxiliary units such as fire weir, THW, police and / or the particularly technical extent of repair of the vehicle that is based at least in part on the identified damage, wherein the estimation algorithm is based on or is based on a data stream oriented programming framework such that by machine learning, particularly by. Arithmetic operations which are carried out by artificial neural networks on multidimensional data fields, so-called tensors, an artificial intelligence (AI) can be generated.According to at least one embodiment, the present invention provides a non-transitory computer readable storage medium, wherein executing the evaluation algorithm comprises executing one or more object detection algorithms that compare one or more images of the vehicle to one or more images of a similar make and model undamaged vehicle.According to at least one embodiment, the method for inspecting a vehicle comprises the following steps:capturing the vehicle at an analysis location of an analysis device, the analysis device comprising at least one camera, for example exactly one camera, directed at the analysis location, the camera configured to generate data characterizing one or more states of the vehicle when the vehicle is positioned at the analysis location, the camera being communicatively connected to a computing device having a memory storing instructions that, when executed, cause the computing device to automatically perform an inspection of the vehicle by executing an inspection algorithm, and receiving the data,the computing device characterizing the one or more conditions of the vehicle; executing, at the computing device of the plurality of cameras; an evaluation algorithm, by the computing device, that causes the computing device to automatically analyze the data and identify damage to the vehicle based on the data;Execution of an estimation algorithm which causes the computing device to automatically perform an estimation of the damage variable for auxiliary units such as fire weir, THW, police and / or the in particular technical scope of the repair of the vehicle, which is based at least in part on the identified damage, wherein the estimation algorithm is based on or is based on a framework for data stream-oriented programming, such that by machine learning, in particular by. Arithmetic operations which are carried out by artificial neural networks on multidimensional data fields, so-called tensors, an artificial intelligence (AI) can be generatedAccording to at least one embodiment, the method further comprises: receiving, by the device, a first damage notification relating to the vehicle from a device configured for automatic detection of vehicle accidents in response to detection of an accident involving the vehicle.The invention is explained in more detail below with reference to a figure.FIG. 1 shows a recording apparatus 100 comprising an analysis device 1 having an analysis location 3 at which a vehicle 10 can be positioned, the analysis device 1 comprising at least one, for example exactly one, sensor 2 directed towards the analysis location 3, wherein sensor 2 is configured to generate sensor data 4 characterizing one or more states of the vehicle 10 when the vehicle 10 is positioned at the analysis location 3, wherein the sensor 2 comprises at least one camera 18, a laser, a mechanical device, a tactile sensor, a liquid sensor, an infrared sensor and combinations thereof; and a computing unit 5 communicatively connected to the sensor 2 and comprising a memory 6 with instructions which, when executed, cause the computing unit 5:initiating execution of an inspection algorithm that automatically performs inspection of the vehicle 10 in response to receipt of the vehicle 10 at the analysis location 3, wherein the computing unit 5 automatically controls operation of the sensor 2 during execution of the inspection algorithm, receiving the sensor data 4 of the sensor 2, wherein the sensor data characterizes the one or more states of the vehicle 10, executing an evaluation algorithm that causes the computing device 5 to automatically analyze the sensor data 4 and automatically identify damage S 1 to the vehicle 10 based on the sensor data 4, and executing an estimation algorithm that causes the computing device 5, The method according to the invention can also be used to automatically estimate the damage variable, inter alia, or also to perform an area expansion on a body surface of the passenger car 10 and / or a depth shift and / or depth change profile of the body surface of the passenger car 10, for auxiliary units such as fire weirs, THWs, police and / or the, in particular, technical extent of the repair of the vehicle 10, which is based at least in part on the identified damage S 1, wherein the estimation algorithm is based on a framework for data stream-oriented programming, so that artificial intelligence (AI) can be generated by machine learning, in particular by arithmetic operations which are carried out by artificial neural networks on multidimensional data fields, so-called tensors.Additionally, FIG. 1 shows a system wherein the memory 6 further causes the computing device 5 to receive, from a device configured to automatically detect vehicle accidents 200, a first damage message related to the vehicle 10 in response to the device 200 detecting an accident involving the vehicle 10 and that the first damage message includes telematics information collected by the device 200 during operation of the vehicle 10.It can be seen from FIG. 1 that the telematics information includes one or more speed measurements, one or more brake measurements, and one or more acceleration measurements of the vehicle 10 either before, after, or before, and after a time of the accident, and that the telematics information includes one or more vehicle parameters collected via an on-board diagnostic system (OBD) 15 connected to the vehicle 10, and that the one or more vehicle parameters include one or more measurements of mileage, coolant temperature, engine speed, spark timing, throttle position, and one or more measurements from an oxygen sensor in the vehicle 10.FIG. 1 also shows a non-transitory computer readable storage medium having stored thereon computer executable program instructions 2000 that, when executed by a processor 16, cause a computing device 17 to (i) initiate execution of an inspection algorithm that automatically performs inspection of a vehicle 10 in response to receipt of data by the sensors 2:(i) initiating execution of an inspection algorithm that automatically performs inspection of a vehicle 10 in response to receipt of the vehicle 10 at an analysis location 3 within an analysis device 1, wherein the analysis device 1 comprises at least one camera 18, for example exactly one camera 18, directed at the analysis location 3, wherein the camera 18 is configured to generate data characterizing one or more states of the vehicle 10 when the vehicle 10 is positioned at the analysis location 3, and wherein the camera 18 is communicatively coupled to the computing device 5 and the computing device 5 automatically controls operation of the camera 1 during execution of the inspection algorithm;(ii) receiving the data characterizing the one or more conditions of the vehicle 10 from the camera 18;(iii) executing an evaluation algorithm that causes the computing device 5 to automatically analyze the data and automatically identify damage S 1 to the vehicle 10 based on the data; and(iv) Execution of an estimation algorithm which causes the computing device 5 to automatically perform an estimation of the damage variable for auxiliary units such as fire weir, THW, police and / or the in particular technical scope of the repair of the vehicle 10, which is based at least in part on the identified damage S 1, wherein the estimation algorithm is based on a framework for data stream-oriented programming, such that by machine learning, in particular by. Arithmetic operations which are carried out by artificial neural networks on multidimensional data fields, so-called tensors, an artificial intelligence (AI) can be generated.Additionally, referring to FIG. 1, a non-transitory computer readable storage medium 2000, wherein executing the evaluation algorithm includes executing one or more object detection algorithms that compare one or more images of the vehicle 10 to one or more images of a similar make and model undamaged vehicle.Reference numeral 6 may be identical to reference numeral 2000.Also, FIG. 1 shows a method for inspecting a vehicle 3000, comprising capturing the vehicle 10 at an analysis location 3 of an analysis device 1, the analysis device 1 comprising at least one camera 18, for example exactly one camera 18, directed towards the analysis location 3, the camera 18 being configured to generate data characterizing one or more states of the vehicle 10 when the vehicle 10 is positioned at the analysis location 3, the camera 18 being communicatively connected to a computing device 5 having a memory storing instructions which, when executed, cause the computing device 5 to automatically perform an inspection of the vehicle 10 by executing an inspection algorithm and receiving the data characterizing the one or more states of the vehicle, in the computing device 5 of the plurality of cameras and executing, by the computing device 5, an evaluation algorithm that causes the computing device 5 to automatically analyze the data and identify damage S 1 to the vehicle 10 based on the data; and executing an estimation algorithm that causes the computing device 5 to automatically make an estimate of the damage magnitude for auxiliary units such as fire weir, THW, police and / or the particularly technical scope of repair of the vehicle 10 based at least in part on the identified damage S 1, wherein the estimation algorithm is based on a data stream oriented programming framework such that by machine learning, particularly by. Arithmetic operations which are carried out by artificial neural networks on multidimensional data fields, so-called tensors, an artificial intelligence (AI) can be generatedFurther, FIG. 1 shows a method 3000 comprising:receiving a first damage message relating to the vehicle 10 from a device 5 configured for the automatic detection of vehicle accidents in response to the detection of an accident involving the vehicle 10 by the device 5.List of reference characters100 Recording apparatus 1 Analysis device 2 Sensor 3 Analysis location 4 Sensor data 5 Computing unit, computing apparatus 6 Memory 10 Vehicle 15 On-board diagnostic system (OBD) 16 Processor 17 Computer apparatus 18 Camera 200 Apparatus for automatically detecting vehicle accidents 2000 Non-transitory computer-readable storage medium having computer-executable program instructions 3000 Method for inspecting a vehicle S 1 Damage stored thereon

Claims

A capturing device (100) comprising: an analysis device (1) having an analysis location (3) at which a vehicle (10) can be positioned, the analysis device (1) comprising at least one, for example exactly one, sensor (2) directed towards the analysis location (3), the sensor (2) being configured to generate sensor data (4) characterizing one or more states of the vehicle (10) when the vehicle (10) is positioned at the analysis location (3), the sensor (2) comprising at least one camera (18), a laser, a mechanical device, a tactile sensor, a liquid sensor, an infrared sensor and combinations thereof; and a computing unit (5) communicatively connected to the sensor (2) and comprising a memory (6) having instructions which, when executed, the computing unit (5) causes: to initiate execution of an inspection algorithm that automatically performs inspection of the vehicle (10) in response to receipt of the vehicle (10) at the analysis site (3), wherein the computing unit (5) automatically controls operation of the sensor (2) during execution of the inspection algorithm, to receive the sensor data (4) of the sensor (2), wherein the sensor data characterizes the one or more states of the vehicle (10), to execute an evaluation algorithm that causes the computing device (5) to automatically analyze the sensor data (4) and to automatically identify damage (S 1) to the vehicle (10) based on the sensor data (4), and to execute an estimation algorithm that causes the computing device (5), An automatic estimation of the damage variable, among other things, thus also or an area expansion on a body surface of the passenger car (10) and / or a depth shift and / or depth change profile of the body surface of the passenger car (10), for auxiliary units such as fire weirs, THWs, police and / or the, in particular, technical scope of the repair of the vehicle (10), which is based at least in part on the identified damage (S1), wherein the estimation algorithm is based on a framework for data stream-oriented programming, so that an artificial intelligence (AI) can be generated by machine learning, in particular by calculation operations which are carried out by artificial neural networks on multidimensional data fields, so-called tensors.The recording device (100) according to claim 1, wherein the memory (6) further causes the computing device (5) to receive, from a device configured to automatically detect vehicle accidents (200), a first damage message related to the vehicle (10) in response to the device (200) detecting an accident involving the vehicle (10).The recording device (100) of claim 2, wherein the first damage message comprises telematics information collected by the device (200) during operation of the vehicle (10).The recording device (100) of claim 3, wherein the telematics information comprises one or more speed measurements, one or more brake measurements, and one or more acceleration measurements of the vehicle (10) either before, after, or before and after a time of the accident.The recording device (100) of claim 3, wherein the telematics information comprises one or more vehicle parameters acquired via an on-board diagnostic system (OBD) (15) connected to the vehicle (10).The recording device (100) of claim 5, wherein the one or more vehicle parameters comprise one or more measurements of mileage, coolant temperature, engine speed, spark timing, throttle position, and one or more measurements from an in-vehicle oxygen sensor (10).A non-transitory computer readable storage medium having stored thereon computer executable program instructions (2000) that, when executed by a processor (16), cause a computing device (17) to (i) initiate execution of an inspection algorithm that automatically performs inspection of a vehicle (10) in response to receipt of data by the sensors (2): (i) initiate execution of an inspection algorithm that automatically performs inspection of a vehicle (10) in response to receipt of the vehicle (10) at an analysis location (3) within an analysis device (1), the analysis device (1) comprising at least one camera (18), for example exactly one camera (18), directed towards the analysis location (3), the camera (18) configured to generate data, characterizing the one or more conditions of the vehicle (10) when the vehicle (10) is positioned at the analysis location (3), and wherein the camera (18) is communicatively coupled to the computing device (5) and the computing device (5) automatically controls operation of the camera (18) during execution of the inspection algorithm; (ii) receiving data characterizing the one or more conditions of the vehicle (10) from the camera (18); (iii) executing an evaluation algorithm that causes the computing device (5) to automatically analyze the data and automatically identify damage (S1) to the vehicle (10) based on the data; (iv) executing an estimation algorithm which causes the computing device (5) to automatically perform an estimation of the damage variable for auxiliary units such as fire weir, THW, police and / or the in particular technical scope of the repair of the vehicle (10), which is based at least in part on the identified damage (S 1), wherein the estimation algorithm is based on a framework for data stream-oriented programming, such that by machine learning, in particular by. Arithmetic operations which are carried out by artificial neural networks on multidimensional data fields, so-called tensors, an artificial intelligence (AI) can be generated.The non-transitory computer readable storage medium (2000) of claim 7, wherein executing the evaluation algorithm comprises executing one or more object detection algorithms that compare one or more images of the vehicle (10) with one or more images of a similar make and model undamaged vehicle.A method of inspecting a vehicle (3000), comprising: capturing the vehicle (10) at an analysis location (3) of an analysis device (1), the analysis device (1) comprising at least one camera (18), for example exactly one camera (18), directed towards the analysis location (3), the camera (18) being configured to generate data characterizing one or more states of the vehicle (10) when the vehicle (10) is positioned at the analysis location (3), the camera (18) being communicatively connected to a computing device (5) having a memory storing instructions that, when executed, cause the computing device (5) to automatically perform an inspection of the vehicle (10) by executing an inspection algorithm, and receiving the data characterizing the one or more states of the vehicle (10), in the computing device (5) of the plurality of cameras (18); executing, by the computing device (5), an evaluation algorithm that causes the computing device (5) to automatically analyze the data and identify damage (S 1) to the vehicle (10) based on the data; and executing an estimation algorithm that causes the computing device (5) to automatically make an estimate of the damage magnitude for auxiliary units such as fire weir, THW, police and / or the particularly technical extent of repair of the vehicle (10) based at least in part on the identified damage (S 1), wherein the estimation algorithm is based on a data stream oriented programming framework such that machine learning, particularly through. Arithmetic operations which are carried out by artificial neural networks on multidimensional data fields, so-called tensors, an artificial intelligence (AI) can be generated.The method (3000) of claim 9, further comprising: receiving, by the device (5), a first damage indication relating to the vehicle (10) from a device (5) configured for automatic detection of vehicle accidents in response to detection of an accident involving the vehicle (10).

Images