Monitoring video real-time compression transmission system and method

Abstract

The embodiment of the application discloses a kind of monitoring video real-time compression transmission system and method, the system includes switch, wireless communication device, camera, DVC device and headgear positioning device, the camera is connected to switch by network cable, switch is configured with the Vlan of corresponding number and monitoring visual angle conversion to camera;DVC device receives the video stream and headgear positioning data of each camera converged from switch, calculates the real-time speed VH of headgear horizontal operation, according to real-time speed VH, the Vlan of video stream and the priority PRI of preset Vlan, different video compression method is used to compress video and upload by wireless communication device.The application effectively reduces the motion estimation time under the premise of ensuring image quality, improves the overall coding performance, improves the video compression efficiency, improves the quality of compressed monitoring video, and solves the problem of large storage space occupied by video raw data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of video compression transmission of crown block, and particularly to a monitoring video real-time compression transmission system and method. BACKGROUND

[0002] With the development of the industrial 4.0 process and the continuous improvement of information technology level, detection and sensor technology, PLC, frequency converter, wireless transmission, video monitoring and other basic automation technologies and information technology are becoming mature, and traditional remote video monitoring technology has gradually begun to be applied in industrial environments.

[0003] In the application scenarios of the steel industry, the video monitoring of unmanned crown block systems has begun to be popularized. A monitoring camera is generally installed on the crown block, and the monitoring video data of the moving crown block is transmitted to the monitoring center's monitoring large screen and NVR (Network Video Recorder) server through 4G / Wifi wireless signals. For example, the utility model patent with application number 2021223034633 discloses a crown block video monitoring system including a video server, a display terminal, multiple groups of cameras installed on the crown block, a switch, a 5G CPE, an RRU, an expansion unit, a baseband processing unit, a 5G core network, etc. The monitoring large screen wants to see high-definition graphics, and the NVR server wants to store video for a longer time. If uncompressed video is directly transmitted to the monitoring center, the transmission pipeline requires high speed and bandwidth, and the storage space of the NVR server also needs to be large enough.

[0004] When transmitting high-definition video, the transmission pipeline requires high speed and bandwidth, and the storage space of the NVR server also needs to be large enough.

[0005] The implementation of the traditional crown block monitoring system is shown in Figure 1 A plurality of cameras are installed on the crown block, the cameras are connected to the switch through a network cable, and the switch is connected to the wireless receiver (wifi AP or 4G receiving device) through a network cable. The wireless receiver transmits data to the wireless transmitter (wifi signal transmitter or base station transmitter) through wireless signals, the wireless transmitter is connected to the router through a network cable or an optical fiber cable, the router is connected to the NVR through a network cable, the video data is transmitted to the NVR, and the display large screen is connected to the router through a network cable to retrieve the real-time data received by the NVR.

[0006] In the above networking scenario, the cameras in the crown block transmit data to the NVR and the display large screen. Since the traditional video is not compressed at the front-end camera, high-definition signals are directly transmitted to the NVR and the display large screen, which puts a large pressure on the entire network, and requires a large storage space of the NVR server. SUMMARY

[0007] The technical problem to be solved by the embodiments of the present application is to provide a monitoring video real-time compression transmission system and method to solve the problem of high definition and small storage space in real-time transmission of crane video in the steel industry.

[0008] To solve the above technical problem, the embodiments of the present application provide a monitoring video real-time compression transmission system, which comprises a switch, a wireless communication device, a camera arranged on a crane at a preset monitoring viewing angle, a DVC device, and a crane positioning device for positioning the position of the crane.

[0009] The DVC device receives the video stream of each camera converged by the switch and the crane positioning data, calculates the real-time speed VH of the horizontal operation of the crane, and compresses the video by using different video compression methods according to the real-time speed VH, the Vlan of the video stream, and the preset priority PRI of the Vlan and uploads the video through the wireless communication device, wherein PRI=0 or 1, 0 represents using a preset first video compression method, and 1 represents using a preset second video compression method.

[0010] Further, the preset first video compression method is an asymmetric cross-shaped multi-level hexagonal grid point search algorithm, and the preset second video compression method calculates a motion vector according to the following steps to eliminate the temporal redundancy of the video signal and improve the coding efficiency.

[0011] Step (1): motion vector prediction is performed according to the H.264 standard algorithm to determine the starting search point; if the SAD value of the starting point is very small, directly jump to step (6); if the SAD value is larger, jump to step (5); only when the SAD value is greater than a determined limit threshold, enter step (2);

[0012] Step (2): calculate the pixel length of the search in the horizontal and vertical directions, and search in the horizontal and vertical directions according to the calculation result by using an asymmetric cross-shaped search; if the SAD value is very small, directly jump to step (6); if the SAD value is larger, jump to step (5);

[0013] Step (3): when VH < preset speed value, enter step (4); when VH ≥ preset speed value, perform spiral search in the region calculated in step (2);

[0014] Step (4): use a multi-heavy hexagonal search mode to gradually reduce the range;

[0015] Step (5): perform multi-turn middle hexagonal template search in the search range;

[0016] Step (6): search with small rhombus as a template to determine the final motion vector.

[0017] Further, in step (2), the pixel length of horizontal and vertical direction search is calculated by the following formula:

[0018] V = W*tanθ i ;

[0019] Wherein, W is the pixel length of horizontal direction search, V is the pixel length of vertical direction search, θ i is the monitoring angle of the camera numbered i.

[0020] Further, the camera is 8 in total, numbered 1~8 respectively, wherein, the cameras 1~4 are installed on the four corners of the crown block respectively, used to monitor the horizontal front and back direction movement; the cameras 5~8 are installed on the middle part of the crown block, monitoring downward, used to monitor the hook below the crown block.

[0021] Further, the crown block positioning device comprises a crown block positioning data device A and a crown block positioning data device B installed on the starting end of the outer guardrail of the crown block and the crown block respectively, the crown block positioning data device A and the crown block positioning data device B are realized through wireless pulse ranging, the real-time speed VH of the horizontal operation of the crown block is (ST i+1 -ST i ) / (T i+1 -T i ), wherein, ST i+1 is the distance between the crown block positioning data device B and the crown block positioning data device A at time T i+1 , and ST i is the distance between the crown block positioning data device B and the crown block positioning data device A at time T i .

[0022] Further, the Vlan value of the video stream of the camera i is i * 100 + θ i ; wherein, θ i is the monitoring angle of the camera numbered i, i is the camera number, and i is an integer.

[0023] Correspondingly, the embodiment of the present application also provides a monitoring video real-time compression transmission method, comprising the following steps:

[0024] S1, number each camera on the crown block, configure the cameras according to the corresponding number and monitoring angle into Vlan through the switch, collect the video stream of each camera and the crown block positioning data;

[0025] S2, calculate the real-time speed VH of the horizontal movement of the crown, and compress the video by using different video compression methods according to the real-time speed VH, the Vlan of the video stream and the preset priority PRI of the Vlan; wherein, PRI=0 or 1, 0 represents using the preset first video compression method, and 1 represents using the preset second video compression method;

[0026] S3, upload the compressed video through the wireless communication device.

[0027] Further, the preset first video compression method is an asymmetric cross-shaped multi-level hexagonal grid search algorithm, and the preset second video compression method calculates a motion vector according to the following steps to eliminate the time redundancy of the video signal and improve the coding efficiency:

[0028] Step S21: perform motion vector prediction according to the H.264 standard algorithm to determine a starting search point; if the SAD value of the starting point is very small, directly jump to step S26; if the SAD value is relatively large, jump to step S25; only when the SAD value is greater than a determined limit threshold, enter step S22;

[0029] Step S22: calculate the pixel length of the horizontal and vertical direction search, and according to the calculation result, use the asymmetric cross-shaped search in the horizontal and vertical direction; if the SAD value is very small, directly jump to step S26; if the SAD value is relatively large, jump to step S25;

[0030] Step S23: when VH < preset speed value, enter step S24; when VH ≥ preset speed value, perform spiral search in the region calculated in step S22;

[0031] Step S24: use a multi-major hexagonal search mode to gradually reduce the range;

[0032] Step S25: perform multi-turn middle hexagonal template search in the search range;

[0033] Step S26: search with a small diamond as a template to determine the final motion vector.

[0034] Further, in step 22, the pixel length of the horizontal and vertical direction search is calculated by the following formula:

[0035] V = W*tanθ i ;

[0036] Wherein, W is the pixel length of the horizontal direction search, V is the pixel length of the vertical direction search, and θ i is the monitoring angle of view of the camera numbered i.

[0037] Further, in step S1, the overhead traveling crane positioning data is acquired by using an overhead traveling crane positioning device, the overhead traveling crane positioning device comprises overhead traveling crane positioning data device A and overhead traveling crane positioning data device B which are respectively installed at a starting end of an outer guardrail of the overhead traveling crane and the overhead traveling crane, the overhead traveling crane positioning data device A and the overhead traveling crane positioning data device B are achieved by wireless pulse ranging, and the real-time horizontal running speed VH of the overhead traveling crane is ST i+1 - ST i ) / (T i+1 - T i ), wherein ST i+1 is the distance between the overhead traveling crane positioning data device B and the overhead traveling crane positioning data device A at time T i , and ST i is the distance between the overhead traveling crane positioning data device B and the overhead traveling crane positioning data device A at time T i .

[0038] The present application has the advantages that the present application processes and compresses the video based on the motion speed of the overhead traveling crane and the monitoring angle of the camera, effectively reduces the motion estimation time, improves the overall coding performance, improves the video compression efficiency, improves the quality of the compressed monitoring video, and solves the problem that the original video data occupies a large storage space. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 is a structural schematic diagram of an existing monitoring video real-time compression transmission system.

[0040] Figure 2 is a structural schematic diagram of a monitoring video real-time compression transmission system of an embodiment of the present application.

[0041] Figure 3 is a top view schematic diagram of a camera installation mode of an embodiment of the present application.

[0042] Figure 4 is a side view schematic diagram of a camera installation mode of an embodiment of the present application.

[0043] Figure 5 is a frame format schematic diagram of a Vlan packet in an Ethernet packet of an embodiment of the present application.

[0044] Figure 6 is a positioning schematic diagram of an overhead traveling crane positioning device of an embodiment of the present application.

[0045] Figure 7 is a search schematic diagram of a preset first video compression method of an embodiment of the present application. EMBODIMENT

[0046] It should be noted that the embodiments and features in the present application can be combined with each other without conflict, and the present application will be further described in detail below in combination with the drawings and specific embodiments.

[0047] In the embodiments of the present application, if there is a directional indication (such as up, down, left, right, front, back, etc.), it is only used to explain the relative position relationship, motion condition, etc. between components in a certain specific posture (as shown in the drawings), and if the specific posture changes, the directional indication will also change accordingly.

[0048] In addition, in the present application, if the description involves "first", "second", etc., it is only for the purpose of description, and cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first" and "second" can explicitly or implicitly include at least one of the features.

[0049] Please refer to Figures 2 to 4 The monitoring video real-time compression transmission system of the embodiment of the present application comprises a switch, a wireless communication device, a camera, a DVC (Digital Video Compressor, video compression processor) device and a crane positioning device.

[0050] The camera has multiple and is numbered in sequence, and is arranged on the crane in a certain monitoring view angle. The camera is connected to the switch through a network cable, and the switch converts the camera according to the corresponding number and monitoring view angle into Vlan for configuration. The crane positioning device adopts a wireless pulse measurement technology to position the crane position, and transmits the position data to the DVC device. The DVC device compresses the camera video. The wireless communication device adopts a 5G private network system, which includes a 5G small base station + 5G core network + 5GCPE, which can significantly increase the bandwidth and transmission rate of the wireless communication pipeline (the 5G small base station adopts an integrated or expanded 5G small base station, the baseband unit of the 5G small base station is connected to the 5G core network through an optical fiber, and the 5G CPE is used to receive the 5G signal of the 5G small base station). The present application improves the transmission speed of the data of the transmission video monitoring through the 5G private network.

[0051] The DVC device receives the video stream of each camera and the crane positioning data converged from the switch, calculates the real-time speed VH of the horizontal operation of the crane, and according to the real-time speed VH, the Vlan of the video stream and the preset priority PRI of the Vlan, adopts different video compression methods to compress the video and upload through the wireless communication device, wherein PRI = 0 or 1, 0 represents using a preset first video compression method, and 1 represents using a preset second video compression method.

[0052] For the camera group with PRI = 0, the motion estimation encoding algorithm of H.264 is completely based on UMHexagonS (Asymmetric Cross Multi-Layer Hexagon Grid Search Algorithm). For the camera with PRI = 1, the motion estimation encoding algorithm of H.264 is based on the second video compression method (i.e. improved UMHexagonS algorithm).

[0053] As an implementation, the preset first video compression method is asymmetric cross multi-layer hexagon grid search algorithm, and the preset second video compression method calculates a motion vector according to the following steps to eliminate time redundancy of a video signal and improve encoding efficiency:

[0054] Step (1): motion vector prediction is performed according to the H.264 standard algorithm to determine a starting search point; if the SAD value of the starting point is very small, directly jump to step (6); if the SAD value is relatively large, jump to step (5); only when the SAD value is greater than a determined limit threshold, enter step (2);

[0055] Step (2): calculate the pixel length of horizontal and vertical direction search, and according to the calculation result, asymmetric cross search is adopted in horizontal and vertical directions; if the SAD value is very small, directly jump to step (6); if the SAD value is relatively large, jump to step (5);

[0056] Step (3): when VH < preset speed value, enter step (4); when VH ≥ preset speed value, perform spiral search in the region calculated in step (2);

[0057] Step (4): adopt a multi-heavy hexagon search mode to gradually narrow the range;

[0058] Step (5): perform multi-circle middle hexagon template search in the search range;

[0059] Step (6): search with a small rhombus as a template to determine the final motion vector.

[0060] As an implementation, in step (2), the pixel length of horizontal and vertical direction search is calculated by the following formula:

[0061] V = W*tanθ i ;

[0062] wherein W is the pixel length of horizontal direction search, V is the pixel length of vertical direction search, θ i is the monitoring angle of view of the camera numbered i.

[0063] As an implementation, the camera is 8, numbered 1~8, wherein the camera 1~4 is installed in the four corners of the crown block, and is used for monitoring the horizontal front and rear direction movement; the camera 5~8 is installed in the middle of the crown block, and is used for monitoring the hook below the crown block.

[0064] As an implementation, the crown block positioning device comprises a crown block positioning data device A and a crown block positioning data device B installed at the starting end of the crown block peripheral guardrail and the crown block respectively, the crown block positioning data device A and the crown block positioning data device B are realized by wireless pulse ranging, and the crown block horizontal operation real-time speed VH = (ST i+1 - ST i ) / (T i+1 - T i ), wherein ST i+1 is the distance between the crown block positioning data device B and the crown block positioning data device A at time T i+1 , and ST i is the distance between the crown block positioning data device B and the crown block positioning data device A at time T i .

[0065] As an implementation, the Vlan value of the video stream of the camera i = i * 100 + θ i ; wherein θ i is the monitoring angle (0~90 degrees) of the camera numbered i, i is the camera number, and i is an integer (the value range of i is determined according to the number of cameras installed on the crown block, and in the case, the number of crown block cameras ranges from 8). Among them, the monitoring angle of the camera 1~4 is the angle between the camera and the horizontal direction, and the monitoring angle of the middle camera 5~8 is the angle between the camera and the vertical direction. In addition, the priority PRI of Vlan can be set by the user, which is used to adopt different algorithms for the camera, 0 represents the default algorithm, and 1 represents the use of a new algorithm.

[0066] The preset second video compression method of the application is an improvement on the UMHexagonS algorithm:

[0067] 1. In the traditional search scheme, when the asymmetric cross search is performed, the horizontal direction is W, and the vertical direction is W / 2. In the application, the horizontal and vertical search values are calculated through the Vlan value corresponding to each camera, and the calculation formula is: V = W*tanθ i , wherein W is the horizontal direction search pixel length (for example, when a 16-pixel length search is adopted, W=16), V is the vertical direction search pixel length, θ i is the downward monitoring angle of the i-th camera, and θ i = Vlan%100. Through the camera monitoring range, the search area is more accurately processed.

[0068] 2. The DVC device calculates the real-time movement speed VH of the overhead crane. When VH < 1 m / s, the UMHexagonS algorithm does not perform a spiral search within the 5×5 region and directly enters the multi-hexagonal template search. When VH => 1 m / s, the original algorithm continues the spiral search within the 5×5 region.

[0069] Its network diagram is shown below. Figure 2 In steel plants, overhead cranes are typically suspended in the air, with hooks underneath for lifting objects. The horizontal travel distance of the crane can reach several hundred meters (the length varies depending on the factory environment, but is generally around 400 meters). Guardrails protect the tracks on both sides of the crane. This invention installs eight cameras on the crane. Cameras 1-4 are installed at the four corners of the crane to monitor its horizontal forward and backward movement. Cameras 5-8 are installed in the middle of the crane, with their lenses facing downwards, to monitor the hooks below. The overhead crane camera installation method, viewed from above, is described in [details omitted]. Figure 3 A side view showing the camera layout of the overhead crane. Figure 4 .

[0070] Overhead crane cameras 1-4 observe the horizontal direction, with an angle generally less than 10 degrees to the horizontal. Cameras 5-8, used to observe hook operation, are installed at a certain angle (generally 30-45 degrees) to the hook (vertical direction) to allow for a more comprehensive view of the hook and remote control. The monitoring angle θ of these cameras... i , represents the monitoring angle of the i-th camera (the angle between the horizontal camera and the horizontal direction, and the angle between the downward monitoring camera and the vertical direction). Figure 4 The monitoring perspectives of camera 1 and camera 5 are given as examples only.

[0071] The overhead crane's cameras are connected to the switch via network cables. To compress the camera data aggregated by the switch, a DVC (video compression processor) is added between the wireless terminal device and the switch. This DVC compresses the received camera video streams before transmitting them through the wireless terminal device. Furthermore, the overhead crane is equipped with a crane positioning data device B. This device B works in conjunction with crane positioning data device A via wireless pulses to transmit distance measurement data and timestamp data to the DVC device through the switch. Crane positioning data device A is installed at one end of the guardrail (the starting point of the crane's operation), and crane positioning data device B is connected to the switch via a network cable.

[0072] After passing through the switch, the overhead crane camera transmits its camera number and monitoring viewpoint information to the DVC device via a Layer 2 VLAN packet encapsulated by the switch. The DVC device then uses this information to perform video compression. This invention transmits the camera number and monitoring viewpoint to the DVC device by setting the switch's VLAN (Virtual LAN) value, enabling the DVC device to perform differentiated video compression encoding. The encoding method is: VLAN value for camera i = i * 100 + θ i Where i is the camera number (i is an integer, determined by the number of cameras installed on the overhead crane; in this embodiment, the number of cameras on the overhead crane ranges from 1 to 8), θ i The monitoring angle of camera i is between 0 and 90 degrees. For example, camera 1 is numbered 1, has a monitoring angle of 10 degrees, and its VLAN = 1*100 + 10 = 110. Camera 5 is numbered 5, has a monitoring angle of 30 degrees, and its VLAN = 5*100 + 30 = 530. Furthermore, the eight cameras in this embodiment can be divided into two categories: one category consists of cameras that monitor the horizontal direction of the overhead crane (…). Figure Three (Cameras 1-4) Another type is the camera that monitors the hook, which monitors the direction of the overhead crane hook, and the monitoring angle is between 0 and 90 degrees. Figure 3 (Cameras 5 through 8). The algorithms used by these cameras for video compression can differ. They are distinguished by their VLAN values. Figure 5 This is a schematic diagram of the frame format of VLAN messages in Ethernet packets. The message is shown in Table 1.

[0073] Table 1

[0074]

[0075] In this embodiment of the invention, the camera device consists of a group of four, forming a 360-degree monitoring effect.

[0076] The second video compression method of this invention is a real-time dynamic video compression method that incorporates the trajectory of the overhead crane. Currently, DVC devices generally use H.264-based compression coding (the coding process includes steps such as intra-frame prediction, inter-frame prediction, residual, transform, quantization, entropy coding, filtering, and inverse transform and inverse quantization). Such devices can effectively compress more than 90% of video data, greatly reducing transmission costs.

[0077] The motion trajectory of the crown block is realized by the crown block positioning data device A and the crown block positioning data device B through wireless pulse ranging. The crown block positioning data device B initiates a request ranging pulse signal according to a certain communication frequency. After the crown block positioning data device A receives the pulse signal of the crown block positioning data device B, a pulse signal is returned. The distance between the two devices can be calculated by multiplying the flight time of the pulse signal between the two devices by the pulse flight speed. The data also contains timestamp information, so the speed of the horizontal movement of the crown block can be obtained.

[0078] The speed formula V = (ST i+1 - ST i ) / (T i+1 - T i ), wherein ST i+1 is the distance between the crown block positioning data device B and the crown block positioning data device A at time T i+1 , and ST i is the distance between the crown block positioning data device B and the crown block positioning data device A at time T i , see Figure 6 .

[0079] Generally, when the crown block moves horizontally, it will be in a uniform low-speed running state, for example: 1~5m / s. Only when the goods need to be lifted and unloaded, the speed will be 0~1m / s.

[0080] In the crown block scene, because there are many cameras and the observed scenes and angles are different, and the speeds of the cameras at different times are also different, a unified compression algorithm cannot flexibly adapt to each camera to make the compression effect optimal. Therefore, the DVC device needs to be able to identify different cameras and support differentiated compression algorithms.

[0081] Currently, the DVC device on the market has multiple channel inputs. In the embodiment of the present application, the video stream accessed by different channels is analyzed to obtain the camera number and monitoring angle through Vlan information. The camera number i = Vlan / 100, the camera monitoring angle θ i = Vlan%100, and whether the camera adopts the improved compression algorithm = the priority PRI value of the Vlan packet (0 indicates not to adopt, and 1 indicates to adopt). The DVC device performs different compression processing according to different numbers of cameras and different angles monitored by the cameras.

[0082] The preset second video compression method of the application is mainly improved based on the UMHexagonS algorithm in H.264, and the preset first video compression method of the application adopts the existing UMHexagonS algorithm (asymmetric cross multi-level hexagonal grid search algorithm), and the original UMHexagonS algorithm process steps are as follows:

[0083] First step: motion vector prediction is carried out according to the H.264 standard algorithm to determine the starting search point (i.e. the search point of the starting frame of the monitoring video). If the SAD (sum of Absolute Difference) value of the starting point is very small, directly jump to the sixth step; if the SAD value is larger, jump to the fifth step; only when the SAD value is greater than the determined limit threshold, the second step is carried out.

[0084] Second step: asymmetric cross search is adopted, and the search horizontal direction length is half of the window width and the window height, as shown in step2 of Figure 7 . If the SAD value is very small, directly jump to the sixth step; if the SAD value is larger, jump to the fifth step.

[0085] Third step: spiral search is carried out in the 5*5 area of the center point obtained in the second step, as shown in setp3-1 of Figure 7 .

[0086] Fourth step: multi-level hexagonal template search, multi-level 16-point hexagonal search mode is adopted, and the range is gradually reduced, as shown in step3-2 of Figure 7 .

[0087] Fifth step: extended hexagonal template search, multi-circle hexagonal template search is carried out in the search range, as shown in step4-1 of Figure 7 .

[0088] Sixth step: small diamond template search with a radius of 1, as shown in step4-2 of Figure 7 . The final motion vector is determined.

[0089] The inventor finds that the above UMHexagonS algorithm is a general optimization processing based on video image search, and for some specific scenes, the effect cannot reach the best. In the process of the crown block running, the motion speed of the crown block and the observation angle of the camera are also important factors that need to be considered in video processing. Therefore, the preset second video compression method of the application is improved based on the above algorithm.

[0090] Improvement point one:

[0091] Asymmetric cross search, the original algorithm generally adopts the vertical search pixel length as half of the horizontal search length, and the asymmetric cross search is based on the camera angle, and the pixel lengths of the horizontal and vertical directions are calculated, V = W*tanθ i Where W is the horizontal direction search pixel length (for example, when 16 pixel length search is adopted, W = 16), V is the vertical direction search pixel length, and θ i is the monitoring angle of view of the i-th camera. Figure Four The monitoring range can be better determined through the length relationship of the two sides in the triangular formula, so that the algorithm can more concentrate on searching the pixel points in the range and improve the search efficiency.

[0092] Embodiment: the camera 5 is installed in the middle of the crown block, monitors the vertical direction, and the angle θ5 between the horizontal direction and the vertical direction is 30 degrees; when the asymmetric cross search is performed and the horizontal search pixel W is 16, V = 16*tan30 = 16 ≈ 16*0.5774 ≈ 9, which belongs to the most concentrated and effective range.

[0093] Second improvement point:

[0094] According to the analysis of the UMHexagonS algorithm characteristics, when the motion speed is low, the image changes little, and it is most likely to find the best matching point in the 5*5 range, that is, the search of Step3-1. When the motion speed is high, the probability of the best matching point near the center is extremely low, and the search of Step3-1 should be skipped, and the search of Step3-2 is directly entered. The industry basically judges the high and low of the motion speed through the motion state calculation formula of the current coding macro block, which is quite complex to process and has lower efficiency. The inventor observes that the speed will have two obvious changes in the running process of the crown block, that is, the speed of the crown block will be kept at 1-5 meters per second during the non-loading and unloading process, and the speed of the crown block will be reduced to 1 meter per second during the loading and unloading process. The speed transmitted by the crown block positioning device can replace the traditional motion state calculation method of the coding macro block, and directly determine whether the corresponding camera executes the search of Step3-1. When the DVC device calculates the real-time motion speed VH of the crown block, when VH < 1 meter / s, the UMHexagonS algorithm does not perform spiral search in the 5*5 region, and directly enters the multiple hexagon template search. When VH >= 1 meter / s, the original algorithm continues to perform spiral search in the 5*5 region.

[0095] The monitoring video real-time compression transmission method of the embodiment of the application comprises the following steps:

[0096] S1, number each camera on the crown block, and configure the camera according to the corresponding number and monitoring view angle into Vlan through the switch, collect the video stream of each camera and the crown positioning data;

[0097] S2, calculate the real-time speed VH of the crown horizontal operation, and compress the video according to the real-time speed VH, the Vlan of the video stream and the preset priority PRI of the Vlan; wherein, PRI=0 or 1, 0 represents using the preset first video compression method, and 1 represents using the preset second video compression method;

[0098] S3, upload the compressed video through the wireless communication device.

[0099] As an embodiment, the preset first video compression method is an asymmetric cross multi-level hexagonal grid search algorithm, and the preset second video compression method calculates the motion vector according to the following steps to eliminate the temporal redundancy of the video signal and improve the coding efficiency:

[0100] Step S21: motion vector prediction is performed according to the H.264 standard algorithm to determine the starting search point; if the starting point SAD value is very small, directly jump to step S26; if the SAD value is larger, jump to step S25; only when the SAD value is greater than the determined limit threshold, enter step S22;

[0101] Step S22: calculate the pixel length of horizontal and vertical direction search, and search the horizontal and vertical direction according to the calculation result; if the SAD value is very small, directly jump to step S26; if the SAD value is larger, jump to step S25;

[0102] Step S23: when VH < preset speed value, enter step S24; when VH ≥ preset speed value, perform spiral search in the area calculated in step S22;

[0103] Step S24: use a multi-major hexagonal search mode to gradually narrow the range;

[0104] Step S25: perform multi-turn middle hexagonal template search in the search range;

[0105] Step S26: search with a small diamond as a template to determine the final motion vector.

[0106] As an embodiment, in step 22, the pixel length of horizontal and vertical direction search is calculated by the following formula:

[0107] V = W*tanθ i ;

[0108] wherein W is the length of the horizontally searched pixels, V is the length of the vertically searched pixels, and θ is the angle of the searched pixels i is the monitoring angle of view of the camera numbered i.

[0109] As an implementation, in step S1, the crown positioning data is acquired by using a crown positioning device, the crown positioning device comprises a crown positioning data device A and a crown positioning data device B respectively installed at the starting end of the crown peripheral guardrail and the crown, the crown positioning data device A and the crown positioning data device B are realized by wireless pulse ranging, and the crown horizontal operation real-time speed VH = (ST i+1 -ST i ) / (T i+1 -T i ), wherein ST i+1 is the distance between the crown positioning data device B and the crown positioning data device A at time T i+1 , and ST i is the distance between the crown positioning data device B and the crown positioning data device A at time T i .

[0110] The application solves the problem of large storage space occupied by video original data and improves the monitoring picture quality by the camera mounting mode, the newly added DVC device and the optimized and improved system algorithm.

[0111] Although the embodiments of the present application have been shown and described, it is to be understood that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A real-time compression and transmission system for surveillance video, comprising a switch, wireless communication equipment, and a camera mounted on an overhead crane at a preset monitoring angle, characterized in that, It also includes DVC equipment and crane positioning equipment for locating the crane's position. There are multiple cameras, which are numbered sequentially. All cameras are connected to a switch via network cables. The switch converts the cameras into VLANs according to their corresponding numbers and monitoring angles for configuration. The DVC device receives video streams from various cameras and crane positioning data aggregated from the switch, calculates the real-time speed VH of the crane's horizontal movement, and compresses the video using different video compression methods based on the real-time speed VH, the VLAN of the video stream, and the preset VLAN priority PRI, and uploads the video through the wireless communication device. Here, PRI=0 or 1, where 0 indicates the use of the preset first video compression method and 1 indicates the use of the preset second video compression method. The preset first video compression method is an asymmetric cross-shaped multi-level hexagonal grid search algorithm, and the preset second video compression method calculates motion vectors according to the following steps to eliminate temporal redundancy in the video signal and improve encoding efficiency: Step (1): Perform motion vector prediction according to the H.264 standard algorithm to determine the starting search point; if the SAD value of the starting point is very small, skip directly to step (6); if the SAD value is large, skip to step (5); only when the SAD value is greater than the determined limit threshold will you proceed to step (2). Step (2): Calculate the pixel lengths for searching in the horizontal and vertical directions. Based on the calculation results, use an asymmetric cross-shaped search in the horizontal and vertical directions. If the SAD value is very small, skip directly to step (6); if the SAD value is large, skip to step (5). Step (3): When VH < preset speed value, proceed to step (4); when VH ≥ preset speed value, perform a spiral search within the area calculated in step (2); Step (4): Use a multi-hexagonal search pattern to gradually narrow down the search area; Step (5): Perform a multi-turn hexagonal template search within the search range; Step (6): Use the small rhombus as a template to perform a search and determine the final motion vector; The VLAN value of the video stream from camera i = i * 100 + θ i ; where θ i Let i be the monitoring view of camera number i, where i is the camera number and can be an integer.

2. The real-time compression and transmission system for surveillance video as described in claim 1, characterized in that, In step (2), the pixel lengths for searching in the horizontal and vertical directions are calculated using the following formula: V = W*tanθ i ; Where W is the horizontal search pixel length, V is the vertical search pixel length, and θ i This refers to the monitoring view of camera number i.

3. The real-time compression and transmission system for surveillance video as described in claim 1, characterized in that, There are a total of 8 cameras, numbered 1 to 8. Cameras 1 to 4 are installed at the four corners of the crane to monitor horizontal forward and backward movement. Cameras 5 to 8 are installed in the middle of the crane, facing downward, to monitor the hooks below the crane.

4. The real-time compression and transmission system for surveillance video as described in claim 1, characterized in that, The crane positioning equipment includes crane positioning data device A and crane positioning data device B, respectively installed at the starting end of the crane's outer guardrail and on the crane itself. Crane positioning data device A and crane positioning data device B are located via wireless pulse ranging. The real-time horizontal speed of the crane is VH = (ST... i+1 - ST i ) / (T i+1 - T i ), among which, ST i+1 For time T i+1 The distance between crane positioning data device B and crane positioning data device A at any given time, ST i For time T i The distance between crane positioning data device B and crane positioning data device A at any given time.

5. A method for real-time compression and transmission of surveillance video, characterized in that, Includes the following steps: S1. Number each camera on the overhead crane, and use a switch to convert the cameras into VLANs according to their corresponding numbers and monitoring angles for configuration. Collect video streams from each camera and overhead crane positioning data. S2. Calculate the real-time speed VH of the crane's horizontal movement. Based on the real-time speed VH, the VLAN of the video stream, and the preset VLAN priority PRI, use different video compression methods to compress the video. Here, PRI = 0 or 1, where 0 indicates the use of the preset first video compression method and 1 indicates the use of the preset second video compression method. S3. Upload the compressed video via wireless communication device; The preset first video compression method is an asymmetric cross-shaped multi-level hexagonal grid search algorithm, and the preset second video compression method calculates motion vectors according to the following steps to eliminate temporal redundancy in the video signal and improve encoding efficiency: Step S21: Perform motion vector prediction according to the H.264 standard algorithm to determine the starting search point; if the SAD value of the starting point is very small, skip directly to step S26; if the SAD value is large, skip to step S25; only when the SAD value is greater than the determined limit threshold will proceed to step S22. Step S22: Calculate the pixel lengths for searching in the horizontal and vertical directions. Based on the calculation results, use an asymmetric cross-shaped search in the horizontal and vertical directions. If the SAD value is very small, skip directly to step S26; if the SAD value is large, skip to step S25. Step S23: When VH < preset speed value, proceed to step S24; when VH ≥ preset speed value, perform a spiral search within the area calculated in step S22. Step S24: Employ a multi-hexagonal search pattern to gradually narrow down the search area; Step S25: Perform a multi-turn hexagonal template search within the search range; Step S26: Use the small rhombus as a template to perform a search and determine the final motion vector; The VLAN value of the video stream from camera i = i * 100 + θ i ; where θ i Let i be the monitoring view of camera number i, where i is the camera number and can be an integer.

6. The real-time compression and transmission method for surveillance video as described in claim 5, characterized in that, In step 22, the pixel lengths for searching in the horizontal and vertical directions are calculated using the following formula: V = W*tanθ i ; Where W is the horizontal search pixel length, V is the vertical search pixel length, and θ i This refers to the monitoring view of camera number i.

7. The real-time compression and transmission method for surveillance video as described in claim 5, characterized in that, In step S1, crane positioning data is acquired using crane positioning equipment. This equipment includes crane positioning data device A and crane positioning data device B, respectively installed at the starting end of the crane's outer guardrail and on the crane itself. Crane positioning data device A and crane positioning data device B are connected via wireless pulse ranging. The real-time horizontal speed of the crane is VH = (ST... i+1 - ST i ) / (T i+1 - T i ), among which, ST i+1 For time T i+1 The distance between crane positioning data device B and crane positioning data device A at any given time, ST i For time T i The distance between crane positioning data device B and crane positioning data device A at any given time.

Images