Special inspection recorder for safety emergency

Abstract

The invention provides a safety emergency special inspection recorder, which comprises a stabilizer, a recorder body, a speed sensor, a local processor and a cloud server, and is characterized in that the recorder body is fixed at a fixed position of a user body through the stabilizer, the speed sensor and the recorder body are both connected with the local processor, and the local processor is remotely connected with the cloud server. And the local processor determines the on-off state of the video stream storage of the recorder body according to the speed state acquired by the speed sensor and the picture change degree of the video stream shot by the recorder body along with time. According to the technical scheme, the directive action is used as a video stream storage starting signal, the local storage size is effectively controlled, and useless information is reduced. And the effective storage ratio of the video stream is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of image communication, and particularly to a special inspection recorder for safety emergency. Background Art

[0002] An inspection recorder is an instrument used to record the work conditions of inspectors. It can record the time when the inspector arrives at a certain location and automatically upload it to the database of the management software to achieve scientific management of the inspection work.

[0003] Inspection recorders are widely used in many fields. For example, in order for staff to better inspect a gas station, a corresponding inspection recorder will be used; for the inspection of general safety hazards, on-site records are often required based on the personnel collection situation and the inspection system.

[0004] However, some common inspection recorders still have certain deficiencies in the process of use. In order to be carried around, the inspection recorder is often small in size, which results in limited storage space. And now the video recording technology is becoming more and more mature year by year, resulting in relatively large video files. But in actual use, in fact, the videos with effective information often account for 20% of the overall video stream. And by using the preset recognition method, based on the characteristics of machine learning, it is often necessary to clarify the situation to be recognized in the training set. However, due to the sudden situations that may occur during the inspection, there are often on-site judgments by the inspectors. Using this method will lead to the omission of inspection information.

[0005] Therefore, it is necessary to provide a special inspection recorder for safety emergency to solve the above technical problems. Summary of the Invention

[0006] The present invention is to solve the problem that in the actual use process of the prior art, in fact, the videos with effective information often account for 20% of the overall video stream. And by using the preset recognition method, based on the characteristics of machine learning, it is often necessary to clarify the situation to be recognized in the training set. However, due to the sudden situations that may occur during the inspection, there are often on-site judgments by the inspectors. Using this method will lead to the omission of inspection information. A special inspection recorder for safety emergency is provided, which solves the above problems by using the user's actions as the video acquisition judgment benchmark.

[0007] The present invention provides a special inspection recorder for safety emergency, which includes a stabilizer, a recorder body, a speed sensor, a local processor and a cloud server. The recorder body is fixed at a fixed position on the user's body through the stabilizer. Both the speed sensor and the recorder body are connected to the local processor, and the local processor is remotely connected to the cloud server. The local processor determines the on-off state of the video stream storage of the recorder body according to the speed state collected by the speed sensor and the degree of change of the video stream captured by the recorder body over time.

[0008] Screen and process the video locally to reduce the processing pressure of the video stream on the cloud and avoid system crashes caused by pressure.

[0009] In a preferred embodiment, the local processor of the special inspection recorder for safety emergency of the present invention includes a local memory, a local buffer, a video stream processor, a mode switch and a transmission module. The local memory is used to temporarily store the recorded video locally and quantitatively transmit it to the cloud server through the transmission module. The local buffer is used to pre-store the video stream segments to be transmitted to the local memory and transfer them to the local memory under the confirmation of the video stream processor. The mode switch is used for the user to switch between the normal storage mode and the automatic storage mode. In the normal storage mode, the local memory stores all the video streams recorded when the recorder body is powered on. In the automatic storage mode, the recorded video stream is stored when the speed sensor detects a stable state after deceleration and the continuous picture change rate of the video stream is less than the first threshold, and the video stream storage is closed when the above situation is exited. The video stream processor is used to identify the video stream and the speed state in the automatic storage mode.

[0010] In a preferred embodiment, the video stream processor of the special inspection recorder for safety emergency of the present invention includes a speed analysis module, a timer, a selection module, a picture analysis module and an output module. The timer is a counter that accurately times according to the frequency and duty cycle of the input pulse signal. The speed analysis module is used to record the speed information, obtain the user's normal moving speed according to the historical information, judge whether the user is in a decelerated state based on the normal moving speed, and transmit the real-time decelerated state signal to the selection module. The selection module is used to continuously receive the video stream and transmit the video stream at the corresponding position to the picture analysis module after receiving the decelerated state signal. The picture analysis module is used to judge the picture change of the received video stream selected by the selection module, and transmit the partial video stream lower than the first threshold to the output module. The output module adds a timestamp to the received video stream according to the pulse signal of the timer and outputs it to the local memory.

[0011] In a preferred embodiment, the specific method for the video stream processor to select videos in the special inspection recorder for safety emergency of the present invention is as follows: S1. The speed analysis module uses the magnitude of the speed with the longest duration after denoising in the historical speeds of the speed sensor as the normal moving speed; S2. The speed analysis module continuously receives the speed information of the speed sensor, and uses the signal that is continuously lower than the normal moving speed received by the speed analysis module as the deceleration state signal, and sends the deceleration state signal and the corresponding start timestamp to the selection module; S3. The selection module transfers and temporarily stores the video stream segments temporarily stored in the local buffer to the local storage module and simultaneously sends them to the frame analysis module; S4. The frame analysis module extracts frames from the video stream after the start timestamp; S5. Determine whether the same proportion of pixels at the same position in each frame of the video stream after frame extraction is lower than the threshold. If so, proceed to step S6; otherwise, delete the video stream after the start timestamp in the local memory; S6. Determine whether one of the translational speeds and deformation ratios of adjacent frame images above the judgment threshold is less than or equal to the translational speed and deformation ratio under the normal moving speed. If so, continuously send the video stream to the local memory until the deceleration state ends; otherwise, delete the video stream after the start timestamp in the local memory.

[0012] Based on the habits of personnel observing items, this technical solution provides a supplementary method for identifying inspected objects on the basis of machine learning of video targets. Different from the active identification technology of inspected objects, this technology determines whether there may be valid information in the captured camera footage based on the habits of the collector, and further supplements the reliability of information identification in the actual video stream when machine learning is not yet mature. It can be used as a reasonable recording scheme before the inspection system completes the machine learning for identifying inspected items, and can also help machine learning obtain a larger information library for accurate training results.

[0013] In a preferred embodiment, the stabilizer of the special inspection recorder for safety emergency of the present invention is a six-axis stabilizer.

[0014] The beneficial effects of the present invention are as follows: This technical solution fully considers the composition of the user's human habits, identifies information by human experience, and is directional collection rather than target collection. It can satisfy the supplementation of important inspection information omitted at the program design level.

[0015] Based on the above principle, this technical solution uses a directional action as the start signal for video stream storage, effectively controls the local storage size, reduces useless information, and ensures the effective storage ratio of the video stream. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of a special inspection recorder for safety emergency; Figure 2 Schematic diagram of the local processor architecture of a special inspection recorder for safety emergency Figure 3 Schematic diagram of the video stream processor architecture of a special inspection recorder for safety emergency Figure 4 Flow chart of video selection of a special inspection recorder for safety emergency

[0017] Reference numerals: 1, stabilizer; 2, recorder body; 3, speed sensor; 4, local processor; 41, local memory; 42, local buffer; 43, video stream processor; 431, speed analysis module; 432, timer; 433, selection module; 434, frame analysis module; 435, output module; 44, mode switch; 45, transmission module; 5, cloud server Detailed implementation manners

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments Embodiment 1

[0019] As Figure 1 shown, a special inspection recorder for safety emergency includes a stabilizer 1, a recorder body 2, a speed sensor 3, a local processor 4 and a cloud server 5. The recorder body 2 is fixed at a fixed position of the user's body through the stabilizer 1. Both the speed sensor 3 and the recorder body 2 are connected to the local processor 4. The local processor 4 and the cloud server 5 are remotely connected. The local processor 4 determines the switch state of the video stream storage of the recorder body 2 according to the speed state collected by the speed sensor 3 and the degree of change of the video stream captured by the recorder body 2 over time. In actual use, in order to avoid acquisition noise caused by the jitter of the speed sensor, in this embodiment, the speed sensor is installed on the stabilizer

[0020] As Figure 2As shown in the figure, the local processor 4 includes a local memory 41, a local buffer 42, a video stream processor 43, a mode switching switch 44, and a transmission module 45. The local memory 41 is used to temporarily store the recorded video locally and quantitatively transmit it to the cloud server 5 through the transmission module 45. The local buffer 42 is used to pre-store the video stream segments to be transmitted to the local memory 41 and transfer them to the local memory 41 when confirmed by the video stream processor 43. The mode switching switch 44 is used for the user to switch between the normal storage mode and the automatic storage mode. In the normal storage mode, the local memory 41 stores all the video streams recorded during the power-on state of the recorder body 2. In the automatic storage mode, when the speed sensor 3 detects a stable state after deceleration and the continuous frame change rate of the video stream is less than the first threshold, the recorded video stream is stored until the above situation is exited and the video stream storage is turned off. The video stream processor 43 is used to identify the video stream and speed state in the automatic storage mode.

[0021] As Figure 3 shown in the figure, the video stream processor 43 includes a speed analysis module 431, a timer 432, a selection module 433, a frame analysis module 434, and an output module 435. The timer 432 is a counter that accurately measures time according to the frequency and duty cycle of the input pulse signal. The speed analysis module 431 is used to record speed information, obtain the user's normal moving speed based on historical information, judge whether the user is in a decelerated state based on the normal moving speed, and transmit the real-time deceleration state signal to the selection module 433. The selection module 433 is used to continuously receive the video stream and transmit the video stream at the corresponding position to the frame analysis module 434 after receiving the deceleration state signal. The frame analysis module 434 is used to judge the frame change of the received video stream selected by the selection module 433 and transmit the partial video stream lower than the first threshold to the output module 435. The output module 435 adds a timestamp to the received video stream according to the pulse signal of the timer 432 and outputs it to the local memory 41.

[0022] As Figure 4 shown in the figure, the specific method for the video stream processor 43 to select videos is as follows: S1. The speed analysis module 431 takes the speed magnitude with the longest continuous duration after denoising in the historical speeds of the speed sensor 3 as the normal moving speed; S2. The speed analysis module 431 continuously receives the speed information of the speed sensor 3, and takes the signal continuously lower than the normal moving speed received by the speed analysis module 431 as the deceleration state signal, and sends the deceleration state signal and the corresponding starting timestamp to the selection module 433; S3. The selection module 433 transmits and temporarily stores the video stream segments temporarily stored in the local buffer 42 to the local storage module and simultaneously transmits them to the frame analysis module 434; S4. The video stream after the starting timestamp is frame-extracted by the screen analysis module 434; S5. Determine whether the same proportion of pixels at the same position in each frame of the video stream after frame extraction is lower than the threshold. If so, proceed to step S6; otherwise, delete the video stream after the starting timestamp in the local memory 41; S6. Determine whether one of the translational speed and deformation ratio of adjacent frame images above the judgment threshold is less than or equal to the translational speed and deformation ratio under the normal moving speed. If so, continuously send the video stream to the local memory 41 until the deceleration state ends; otherwise, delete the video stream after the starting timestamp in the local memory 41.

[0023] Optionally, the stabilizer 1 is a six-axis stabilizer 1.

[0024] This system physically compensates the camera through the stabilizer to prevent strong jitter during video recording and at the same time cause large fluctuations in the collection of speed sensor parameters, resulting in affecting the judgment of actual parameters.

[0025] When the user's attention is concentrated, they often slow down to observe to prevent other accidents when their observation ability is insufficient. At the same time, for important scenes, the observer often keeps their attention at a fixed position, which actually leads to the fact that scenes with relatively large information are often directional from the user's perspective. When the conditions of deceleration and fixed observation point are met, it is often possible to draw a conclusion that the observed object has certain particularities.

[0026] During the inspection process, two modes are often used to control local storage.

[0027] One is the user's active switch. The objectivity of this operation mode depends on the operator himself. When the operator selectively records, it often leads to a reduction in the objectivity of the video. The second is to use machine learning to pre-specify the objects to be inspected in advance, which is also restricted by the types of problems recognized in the machine learning training set and is difficult to reflect the recording situations that may be caused by on-site emergencies.

[0028] This embodiment adopts a third method, which conducts inspection records based on the operator's instinctive habitual behavior, and at the same time avoids the operator's actual intervention in the recording selection, ensuring the objectivity of the inspection video.

[0029] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. A special inspection recorder for safety emergency, characterized by: The device comprises a stabilizer (1), a recorder body (2), a speed sensor (3), a local processor (4) and a cloud server (5); the recorder body (2) is fixed to a fixed position on the user's body through the stabilizer (1); the speed sensor (3) and the recorder body (2) are both connected to the local processor (4); the local processor (4) and the cloud server (5) are remotely connected; the local processor (4) determines the on / off state of the video stream storage of the recorder body (2) according to the speed state collected by the speed sensor (3) and the degree of image change over time of the video stream captured by the recorder body (2).

2. A safety emergency inspection recorder according to claim 1, characterized in that: The local processor (4) comprises a local memory (41), a local buffer (42), a video stream processor (43), a mode switching switch (44) and a transmission module (45). The local memory (41) is used to temporarily store recorded videos locally and transmit them to the cloud server (5) in a quantitative manner through the transmission module (45). The local buffer (42) is used to pre-store video stream segments to be transmitted to the local memory (41) and transfer them to the local memory (41) upon confirmation by the video stream processor (43). The mode switching switch (44) is used by a user to switch between a conventional storage mode and an automatic storage mode. In the conventional storage mode, the local memory (41) stores all video streams recorded when the recorder body (2) is powered on. In the automatic storage mode, when the speed sensor (3) detects a stable state after deceleration and the continuous picture change rate of the video stream is less than a first threshold, the recorded video stream is stored until the video stream storage is turned off when the above situation is eliminated. The video stream processor (43) is used to identify the video stream and speed state in the automatic storage mode.

3. A safety emergency inspection recorder according to claim 2, characterized in that: The video stream processor (43) comprises a speed analysis module (431), a timer (432), a selection module (433), a picture analysis module (434) and an output module (435). The timer (432) is a counter that accurately measures time according to the frequency and duty cycle of an input pulse signal. The speed analysis module (431) is used to record speed information and obtain a user's normal moving speed according to historical information, and judge whether the user is in a deceleration state based on the normal moving speed, and transmit a real-time deceleration state signal to the selection module (433). The selection module (433) is used to continuously receive a video stream and transmit a video stream at a corresponding position to the picture analysis module (434) after receiving the deceleration state signal. The picture analysis module (434) is used to perform picture change judgment on the video stream selected by the selection module (433) and transmit a portion of the video stream below a first threshold to the output module (435). The output module (435) adds a timestamp to the received video stream according to the pulse signal of the timer (432), and outputs the value to the local memory (41).

4. A safety emergency inspection recorder according to claim 3, characterized in that: The specific method of the video stream processor (43) selecting a video is: S1, the speed analysis module (431) uses the speed with the longest duration after denoising in the historical speed of the speed sensor (3) as the normal moving speed; S2, the speed analysis module (431) continuously receives speed information from the speed sensor (3), and uses the signal received by the speed analysis module (431) that is continuously lower than the normal moving speed as a deceleration state signal, and sends the deceleration state signal and the corresponding starting point timestamp to the selection module (433); S3, the selection module (433) transmits the video stream segments temporarily stored in the local buffer (42) to the local storage module and temporarily stores them to the picture analysis module (434); S4, the picture analysis module (434) extracts frames from the video stream after the starting time stamp; S5, determining whether the ratio of pixels at the same position in each frame of the video stream after frame extraction is lower than a threshold, if yes, proceeding to step S6, otherwise deleting the video stream after the starting time stamp in the local memory (41); S6. Determine whether one of the translation speed and deformation ratio of adjacent frames above the judgment threshold is less than or equal to the translation speed and deformation ratio at the normal moving speed. If so, continue to send the video stream to the local memory (41) until the deceleration state ends. Otherwise, delete the video stream after the starting time stamp in the local memory (41).

5. A safety emergency inspection recorder according to claim 1, characterized in that: The stabilizer (1) is a six-axis stabilizer.

Images