阅读说明：本技术 一种基于高速激光雷达的铁路货运几何超限预警系统 (Railway freight transportation geometric overrun early warning system based on high-speed laser radar ) 是由 宁星 蓝贤桂 赖毅辉 蓝志鹏 肖佳珠 郭江涛 于 2021-08-16 设计创作，主要内容包括：本发明属于铁路货运领域,具体涉及一种基于高速激光雷达的铁路货运几何超限预警系统,包括设置在室外的前端数据采集子系统和设置在室内的机房子系统,前端数据采集子系统和机房子系统之间通过网络信息交互；采用激光雷达实现了列车高速通过时对车厢外轮廓尺寸的动态检测,不影响正常行车；根据国家铁总公司标准自动判定机车车辆左侧、右侧、顶部三部位超限和车门未关等异常情况,并对异常部位进行定位；以单节车厢为单位自动生成每节车辆左侧、右侧、顶部轮廓图,根据轮廓图,检车员可以快速判断超限具体位置,有效排除超限安全隐患。(The invention belongs to the field of railway freight, and particularly relates to a railway freight geometric overrun early warning system based on a high-speed laser radar, which comprises a front-end data acquisition subsystem arranged outdoors and a machine room subsystem arranged indoors, wherein the front-end data acquisition subsystem and the machine room subsystem are interacted through network information; the laser radar is adopted to realize the dynamic detection of the outer contour size of the carriage when the train passes through at a high speed, and the normal running is not influenced; automatically judging abnormal conditions such as the overrun of three parts of the left side, the right side and the top of the rolling stock, the unclosed door and the like according to the national iron headquarters standard, and positioning the abnormal parts; the left side, the right side and the top contour map of each vehicle are automatically generated by taking a single carriage as a unit, and according to the contour maps, a vehicle inspector can quickly judge the specific position exceeding the limit, so that the potential safety hazard exceeding the limit is effectively eliminated.)

1. A railway freight transportation geometric overrun early warning system based on high-speed laser radar is characterized by comprising a front-end data acquisition subsystem arranged outdoors and a machine room subsystem arranged indoors, wherein the front-end data acquisition subsystem and the machine room subsystem are interacted through network information;

the front-end data acquisition subsystem comprises: the AEI train number recognition device 2 is arranged in the train number recognition system, and is used for automatically recognizing the train number of the running train and uploading the train number to the machine room subsystem;

a vehicle passing detection sensor 1 is arranged in the vehicle passing detection and carriage division unit, the time of a single vehicle entering and leaving the front-end data acquisition subsystem is recorded, and the time is uploaded to the machine room subsystem;

the laser radar group comprises a left laser radar 3 and a right laser radar 4, performs vehicle appearance scanning and overrun detection, and uploads the vehicle appearance scanning and overrun detection to the machine room subsystem;

the machine room subsystem comprises: the system comprises an overrun detection processing host, a vehicle detection application platform and a data storage server.

2. The railway freight transportation geometric transfinite early warning system based on the high-speed laser radar as claimed in claim 1, wherein the laser radar group is arranged on a portal frame outside a track, the left laser radar 3 and the right laser radar 4 are arranged at the same height on the left side and the right side of the portal frame, and laser emitting ends of the left laser radar 3 and the right laser radar 4 are not opposite.

3. The railway freight transportation geometric overrun early warning system based on the high-speed laser radar as claimed in claim 1, wherein the laser radar is a two-dimensional photoelectric measurement system.

4. The geometric transfinite early warning system for railway freight transportation based on high-speed laser radar as claimed in claim 1, wherein the passing detection sensor 1 of the passing detection and carriage division unit is a wheel sensor, and is installed on the lower side of the inner wall of the track for determining whether a train is passing, and information such as the direction of the train passing, the number of axles of the wheels, the running speed of the train, the time of the wheel passing sensor, and the like.

5. The railway freight transportation geometric transfinite early warning system based on the high-speed laser radar as claimed in claim 1, wherein the AEI car number recognition device 2 is arranged in the car number recognition system, the AEI car number recognition device is arranged on the track, the AEI car number recognition device is parallel to a passing train, a power amplification opening instruction is sent through an upper computer, a radio frequency module starts to emit radio frequency signals to the periphery, and tag information at the bottom of the train is recognized and read in this way.

6. The geometric transfinite early warning system for railway freight transportation based on the high-speed laser radar as claimed in claim 1, wherein the measuring and mapping of the outline of the freight vehicle by the transfinite detection processing host machine comprises: setting a geometric overrun judgment standard, preprocessing laser radar data, extracting laser radar point cloud key points, registering point cloud, measuring the outer contour dimension of the rolling stock and mapping the point cloud.

7. The railway freight transportation geometric overrun early warning system based on the high-speed laser radar as claimed in claim 1, wherein the vehicle detection application platform is a geometric overrun detection application software platform;

automatically acquiring information such as train passing number, train type, train speed, passing time and the like of a train;

acquiring and processing the data of the outline dimension of the railway wagon, completing data segmentation by taking a carriage as a unit, and automatically generating a left side view, a right side view and a top view of the outline of the train;

and the inspection personnel realize the vehicle inspection operation through client software.

8. The railway freight transportation geometric overrun early warning system based on the high-speed laser radar as claimed in claim 7, wherein the communication mode of the geometric overrun detection application software platform is Socket communication, and the Socket communication is used for information data interaction between a front-end railway vehicle number identification and detection system, a vehicle passing detection and carriage segmentation system and a Web server database, and is an interface between a user process and a TCP/IP protocol.

9. The geometric transfinite early warning system for railway freight transportation based on the high-speed laser radar as claimed in claim 8, wherein the system database of the data storage server is provided with three database relation tables including a passing car basic information table, a car detection information table and a user information table.

10. The railway freight transportation geometric overrun early warning system based on the high-speed laser radar as claimed in claim 1, wherein client software of the geometric overrun detection application software platform is arranged on a computer device, the computer device comprises a processor and a memory, and the memory is coupled with the processor and used for storing the client software; the system comprises a user login module, a vehicle detection list module, an overrun state vehicle detection main interface, a report module and a user setting module, wherein the user login module, the vehicle detection list module, the overrun state vehicle detection main interface, the report module and the user setting module are interacted with computer equipment.

Technical Field

The invention relates to the field of railway freight, in particular to a railway freight geometric overrun early warning system based on a high-speed laser radar.

Background

In order to ensure the safety of railway freight transportation, the national railway general company strictly regulates the freight train loading limit and the transportation of over-limit cargos, the railway regulates various special limits, and the common cargo over-limit comprises ultrahigh and ultra-wide cargos and the like. At present, when the loading of a railway in China detects the overload of a truck, the truck owner is manually and outdoors finished in a static state, a cargo inspector generally uses a marker post or a sliding ruler to measure the loading width and height after the truck is stopped and is even visually checked by experience, the accuracy is poor, the measurement precision is not high, the efficiency is low, the labor intensity is high, the omission ratio is high, the problems that the running truck cannot be monitored and the like are caused, and safety accidents caused by the overload of the loading occur, at present, systematic research and practice have been carried out on the conditions of superelevation, superwidth and the like at home and abroad, and the common methods comprise a CCD mode identification technology, a laser wheel rotation technology, a fixed point infrared technology and the like, and are mainly divided into the following 2 types:

1. laser/infrared correlation detection mode, mainly according to the boundary limit judgement standard, adopt transmitting terminal and receiving terminal correlation mode to install, set up laser or infrared light beam along the boundary limit direction, carry out "carving the limit" and detect, when the locomotive transfinites, the transfinite position can shelter from the signal production, forms transfinite alarm signal, this type has more meticulous discrimination ability, generally can detect the transfinite more than diameter 3mm, can carry out dynamic verification to the train in marcing, the problem mainly has: only judging whether the overrun exists or not, and being incapable of measuring specific overrun values and specific overrun parts; the actual detection situation on site cannot be reserved, and recording, playback and analysis cannot be realized; environmental interference is very likely to cause false alarms.

2. Image identification and measurement detection mode adopt the camera to carry out video acquisition through the train to carry out image processing and discernment, realize loading the dynamic on-line measuring of transfiniting, this mode can fix a position the concrete position of transfiniting, and data file and inquiry are simple, and the problem exists mainly: the image processing data volume is large, the real-time performance is poor, and the requirement on the performance of an image processor is high; the method is susceptible to environmental interference to cause false alarm; the complex climate and night tend to result in poor image quality.

Disclosure of Invention

Objects of the invention

In order to solve the technical problems in the background technology, the invention provides a railway freight transportation geometric overrun early warning system based on a high-speed laser radar, which has high resolution and can detect overrun of a part with the diameter of more than 5cm under the speed of 20 km/h; the reliability is high, and the overrun abnormality can be effectively detected in the daytime and at night; the real-time performance is good, and the processing of the overrun data can be completed within 5 seconds after the train passes through; strong environmental adaptability and can be used in various climatic environments.

(II) technical scheme

In order to solve the technical problems, the invention provides a railway freight transportation geometric overrun early warning system based on a high-speed laser radar, which comprises a front-end data acquisition subsystem arranged outdoors and a machine room subsystem arranged indoors, wherein the front-end data acquisition subsystem and the machine room subsystem are interacted through network information;

the front-end data acquisition subsystem comprises: the AEI train number recognition device is arranged in the train number recognition system, and is used for automatically recognizing the train number of the running train and uploading the train number to the machine room subsystem;

a vehicle passing detection sensor is arranged in the vehicle passing detection and carriage division unit, the time of a single vehicle entering and leaving the front-end data acquisition subsystem is recorded, and the time is uploaded to the machine room subsystem;

the laser radar group comprises a left laser radar and a right laser radar, performs vehicle appearance scanning and overrun detection, and uploads the vehicle appearance scanning and overrun detection to the machine room subsystem;

the machine room subsystem comprises: the system comprises an overrun detection processing host, a vehicle detection application platform and a data storage server.

Preferably, the laser radar group is arranged on a portal frame outside the track, the left laser radar and the right laser radar are arranged at the same height on the left side and the right side of the portal frame, and the laser emitting ends of the left laser radar and the right laser radar are not opposite.

Preferably, the lidar is a two-dimensional photoelectric measurement system.

Preferably, the passing detection sensor of the passing detection and car division unit is a wheel sensor, which is mounted on the lower side of the inner wall of the track and used for determining whether a train is passing, the direction of the train when passing, the number of axles of the wheels, the train running speed, the time when the wheel passes through the sensor, and other information.

Preferably, the AEI car number recognition device is arranged in the car number recognition system and is installed on the track, the AEI car number recognition device is parallel to a passing train, a power amplifier opening instruction is sent through the upper computer, the radio frequency module starts to emit radio frequency signals to the periphery, and tag information at the bottom of the train is recognized and read in the mode.

Preferably, the measuring and mapping of the contour of the freight vehicle by the overrun detection processing host computer comprises: setting a geometric overrun judgment standard, preprocessing laser radar data, extracting laser radar point cloud key points, registering point cloud, measuring the outer contour dimension of the rolling stock and mapping the point cloud.

Preferably, the vehicle detection application platform is a geometric overrun detection application software platform;

automatically acquiring information such as train passing number, train type, train speed, passing time and the like of a train;

acquiring and processing the data of the outline dimension of the railway wagon, completing data segmentation by taking a carriage as a unit, and automatically generating a left side view, a right side view and a top view of the outline of the train;

and the inspection personnel realize the vehicle inspection operation through client software.

Preferably, the communication mode of the geometry overrun detection application software platform is Socket communication, and the Socket communication is used for information data interaction between a front-end railway vehicle number identification detection system, vehicle passing detection and carriage segmentation system and a Web server database, and is an interface between a user process and a TCP/IP protocol.

Preferably, the system database of the data storage server is provided with three database relation tables including a passing car basic information table, a car checking information table and a user information table.

Preferably, the client software of the geometry overrun detection application software platform is arranged on a computer device, and the computer device comprises a processor and a memory, wherein the memory is coupled with the processor and used for storing the client software; the system comprises a user login module, a vehicle detection list module, an overrun state vehicle detection main interface, a report module and a user setting module, wherein the user login module, the vehicle detection list module, the overrun state vehicle detection main interface, the report module and the user setting module are interacted with computer equipment.

The technical scheme of the invention has the following beneficial technical effects:

1. the laser radar is adopted to realize the dynamic detection of the outer contour size of the carriage when the train passes through at a high speed, and the normal running is not influenced;

2. automatically judging abnormal conditions such as the overrun of the left side, the right side and the top 3 part of the rolling stock, the unclosed door and the like according to the national iron headquarters standard, and positioning the abnormal parts;

3. the left side, right side and top contour diagrams of each vehicle are automatically generated by taking a single carriage as a unit, and according to the contour diagrams, a vehicle inspector can quickly judge the specific position of overrun, so that the potential safety hazard of overrun is effectively eliminated;

the railway freight train geometric overrun early warning system utilizes a high-speed laser radar scanning technology to collect the external contour dimension data of each train in real time and send the external contour dimension data to the overrun detection processor, so that the labor burden of workers is reduced, the working efficiency is improved, and the dynamic overrun detection of the railway freight train without stopping is realized.

Drawings

FIG. 1 is a schematic diagram of a geometric overrun warning system for railway freight in accordance with the present invention;

FIG. 2 is a schematic diagram of a radar scanning architecture of the present invention;

FIG. 3 is a flow chart of point cloud registration of the present invention;

FIG. 4 is a flow chart of a vehicle passing detection and compartment segmentation unit system;

FIG. 5 is a flow chart of a railway car number identification system based on RF-ID;

FIG. 6 is a diagram of a laser radar data distribution when a vehicle passes through a laser radar sensor;

FIG. 7 is a laser radar data distribution diagram when an organic vehicle passes through;

FIG. 8 is a diagram of a filtered lidar data distribution;

FIG. 9 is a cloud point view of the outer contour of the rolling stock 3;

FIG. 10 is a diagram of a system for detecting locomotive overrun

Fig. 11 is a car inspection interface diagram of a car inspector.

Reference numerals: 1. a vehicle passing detection sensor; 2. an AEI car number identification device; 3. a left-side laser radar; 4. and (4) a laser radar on the right side.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings in conjunction with the following detailed description. It should be understood that the description is intended to be exemplary only, and is not intended to limit the scope of the present invention. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present invention.

As shown in fig. 1-5, the railway freight transportation geometric overrun early warning system based on the high-speed laser radar realizes the following main functions:

1. a point cloud data acquisition and processing function;

2. a point cloud image service function and an information integration query function;

3. the functions are based on the cooperation of a front-end data acquisition subsystem arranged outdoors and a machine room subsystem arranged indoors, wherein the front-end data acquisition subsystem and the machine room subsystem are interacted through network information.

The front-end data acquisition subsystem comprises:

the AEI train number recognition device 2 is arranged in the train number recognition system, and is used for automatically recognizing the train number of the running train and uploading the train number to the machine room subsystem;

a vehicle passing detection sensor 1 is arranged in the vehicle passing detection and carriage division unit, the time of a single vehicle entering and leaving the front-end data acquisition subsystem is recorded, and the time is uploaded to the machine room subsystem;

laser radar group carries out vehicle outward appearance scanning and transfinite detection including installing left side laser radar 3 and the right side laser radar 4 in the track both sides to upload to the computer lab subsystem.

The computer room subsystem includes: overrun detects processing host computer, examines car application platform and data storage server, and wherein overrun detects processing host computer and mainly accomplishes:

1. collecting laser radar data and filtering invalid data;

2. processing laser radar data;

3. receiving related data uploaded by an AEI vehicle number identification host, a vehicle passing detection unit and a carriage segmentation unit;

4. the data are correlated, and the correlated data are transmitted to a vehicle detection application platform and a data storage server; thereby realizing the point cloud data acquisition and processing function. The laser radar is a two-dimensional photoelectric measuring system, wherein one dimension is a laser ranging system which measures the distance between the laser radar and the surrounding environment by using laser, the other dimension is an angle control system which controls the angle between the time-measuring laser and the surrounding environment by controlling a motor, the scanning surface of the surrounding environment is determined by a distance measuring value and an angle value, the laser radar can realize the dynamic measurement of parameters such as the distance, the direction, the speed, the shape and the like of a detected target by emitting laser beams, can be applied to the field of railway overrun detection and realizes the collection and the mapping of train body outline information, compared with a manual outdoor field detection mode, the dynamic overrun detection of a passing train can be realized by adopting the laser radar, the detection mode is changed from outdoor to indoor, the train detection efficiency is improved, the labor intensity is reduced, compared with a laser/infrared correlation detection mode and 2 automatic dynamic monitoring modes of image identification and measurement detection, the laser radar has the following advantages: the overrun part can be accurately positioned; is not easy to be interfered by the environment; the device is not easily influenced by the climate environment and dark light at night; the data volume is small, the real-time performance is good, and a railway overrun detection system is designed by utilizing the laser radar.

Wherein the parameters of the laser radar are as follows: the scanning frequency of the laser radar is 200Hz, the angular resolution is 1 degree, the working area is 0.5m … 25m @10 percent, the scanning angle is 110 degrees, the light spot divergence angle is 6mrad, the light source wavelength is 905nm, the adaptive vehicle speed is less than or equal to 50km/h, the detection precision-width is less than or equal to 150mm, the detection precision-height is less than or equal to 50mm, and the Ethernet TCP/IP is 100 Base-TX.

Specifically, the passing detection sensor 1 of the passing detection and compartment division unit is a wheel sensor and is installed on the lower side of the inner wall of the track, when the magnetic field intensity of the internal permanent magnet changes due to the passing of train wheels, the magnetic flux of the induction coil changes, the induction coil instantaneously generates disturbance induced electromotive force to form induced electromotive force, and the induced electromotive force formed by the induction coil is amplified by the voltage amplifier and output to the signal processing module for judging whether a train passes through and the direction of the train when the train passes through, the number of axles of the wheels, the train running speed, the time when the wheels pass through the sensor and other information.

Specifically, an AEI train number recognition device 2 is arranged in the train number recognition system and is installed on a track, the AEI train number recognition device is parallel to a passing train, a power amplifier opening instruction is sent through an upper computer, a radio frequency module starts to emit radio frequency signals to the periphery, tag information at the bottom of the train is recognized and read in such a mode, the read train number information is uploaded to a server database, and each train carriage with train number information is matched with a corresponding train number by combining a train passing detection and carriage division system.

In the embodiment, the laser radar set is arranged on a portal frame outside a track, the left laser radar 3 and the right laser radar 4 are arranged at the same height on the left side and the right side of the portal frame, the laser emitting ends of the left laser radar 3 and the right laser radar 4 are not opposite, the distance between the two laser radars completely covers the track to ensure that a vehicle drives into a measuring area, the lane is assumed to be regular and flat, the vehicle to be measured drives straight into the measuring area (straight can be understood as that the vehicle driving direction does not incline relative to the lane), the vehicle driving direction is set as a positive z-axis direction, a three-dimensional rectangular coordinate system is formed by combining with a laser radar measuring coordinate system, an xOz plane of the laser radar measuring coordinate system needs to be parallel to the road surface, the xOy plane needs to be perpendicular to the road surface and has an included angle of 90 degrees with the vehicle driving direction, and the laser emitting parts of the two radars can not be opposite, otherwise, the interference is generated, and meanwhile, the wheel sensor can measure the vehicle speed, so that the measuring system can acquire data correctly.

It needs to be supplemented that after the hardware system of the measuring system is built, the system calibration work needs to be carried out firstly, the measuring coordinate systems of the two laser radars are registered to the same coordinate system, then the system enters a normal working state, when the train passes through a measuring area, the measuring system obtains the profile data of the train through the laser radars, and the running speed and the running direction of the train are obtained through the wheel sensors.

The work flow of dynamic detection of the railway freight overrun state is as follows:

step one, identifying information such as the running direction and speed of a train by a wheel sensor when the train passes by, and dividing a carriage; the method comprises the following steps that when a train passes through an AEI train number recognition device, an electronic tag at the bottom of a carriage is recognized, carriage information is obtained and transmitted to a server; the high-speed laser radar collects the outline size data of each vehicle in real time and sends the outline size data to the machine room subsystem;

secondly, the overrun detection processing host machine carries out laser radar data acquisition, invalid data filtering and laser radar data processing; receiving related data uploaded by an AEI vehicle number identification host, a vehicle passing detection unit and a carriage segmentation unit; and correlating various data, and transmitting the correlated data to a vehicle detection application platform and a data storage server.

Specifically, the method comprises the following steps: when a train passes through the laser radar overrun detection acquisition area, the train passing detection sensor detects a first wheel signal of the train, the laser radars on the left side and the right side are triggered to start data acquisition, and the acquired data contain invalid environmental background data, so that the invalid data need to be filtered.

Wherein, the laser radar data collection is that when the train passes through the laser radar collection area, the laser radar starts to send laser beams to the surface of the detected train body, a reflected signal is formed on the surface of the train body, a laser radar receiver receives the reflected signal, the distance of the laser radar reaching the detected object is calculated according to the difference between the emitting time and the receiving time of the same laser beam, and the distance is compared with the over-limit standard to judge whether the over-limit exists or not, when the locomotive vehicle passes through, the laser radar is started to emit the laser beams, the reflected signal is simultaneously received to obtain distance data, each distance value is equivalent to a scanning point, the collection of the points is called point cloud, the position of the generated scanning point is shown as a dotted line position in figure 2, the distance between adjacent sections can be known according to the real-time speed information provided by the vehicle passing detection sensor, and the complete three-dimensional outer contour point cloud data of the whole carriage can be obtained after the whole carriage passes through, carrying out locomotive vehicle overrun detection according to the point cloud data;

the point cloud data acquired by the laser radar needs to be subjected to invalid data filtering:

after the laser radar is installed in an overrun detection acquisition area, the acquired point cloud data mainly comprises 2 types of environment background point cloud data and rolling stock outer contour point cloud data, overrun detection is to judge whether overrun exists according to the rolling stock outer contour point cloud data, the environment background point cloud data are invalid data, in order to eliminate the influence of the invalid data on the overrun detection, the environment background point cloud data must be filtered, and the filtering process is as follows:

as shown in fig. 6, when no rolling stock passes through, the laser radars on both sides mainly collect the environmental background point cloud data, and according to the rolling stock size standard and the laser radar installation position, the effective area of the point cloud data can be calculated according to the distance, so as to obtain the laser radar collected data distribution map when the rolling stock passes through, wherein the effective area is the middle gray part, and the point cloud data on the outer side represents the data collected by the laser radars.

As shown in fig. 7, the laser radar collects a data distribution map when the locomotive passes through, wherein the portion in the gray area is the point cloud data of the outer contour of the locomotive.

As shown in fig. 8, when the vehicle passes through, the gray area boundary is used as a point cloud data filtering condition, the collected data is filtered, and only the point cloud data in the gray area is retained, so that a large amount of invalid data is filtered out, valid data is retained, and a filtered laser radar collected data distribution map is obtained.

The following are used as supplements: the filtering processing used for filtering the point cloud data is operation on certain specific frequency in a known signal, and can reduce or avoid external interference to a certain extent, although data collected by a two-dimensional laser radar is relatively pure, the laser radar can generate certain noise, after isolated points are removed, noise still exists in data of each frame of laser radar, unreliable laser reflection caused by background noise can also cause more or less influence on an obstacle detection effect, so that further filtering processing is adopted on the collected laser radar data, the average filtering is a continuous window filtering algorithm, additional new data can be added after filtering, the data cause pseudo obstacle information, and the average filtering is suitable for continuous data columns with not particularly large change amplitude.

Specifically, the point cloud data acquired by the laser radar for the rolling stock is a data set of multiple measurements of the laser radar, each measurement is not performed in the same coordinate system, the point cloud data obtained by the multiple measurements must be unified in the same coordinate system, and massive point cloud data needs to be converted into an outer contour image of the rolling stock, so that the laser radar data needs to be further processed, and the processing process comprises point cloud registration and point cloud mapping.

The point cloud registration process comprises the following steps:

1. point cloud data reading, namely respectively reading point cloud data collected by the left laser radar and the right laser radar;

2. extracting laser radar data at the top of the rolling stock, namely extracting 2 laser radar rolling stock top data from the point cloud data, wherein the collected top point cloud data can be regarded as being on one plane under general conditions, so that the rolling stock top point cloud data are segmented by adopting a plane segmentation method, and the rolling stock top point cloud data are respectively extracted by utilizing a geometric model segmentation extraction module based on a random sampling consistency algorithm in a point cloud library;

3. the method comprises the following steps of (1) carrying out manual coarse registration, wherein a basic coordinate conversion method is utilized for the manual coarse registration, firstly, the position deviation of 2 laser radars is estimated according to the installation positions of the 2 laser radars, the distance between two pieces of top point cloud data of the 2 laser radars is shortened, a good initial value is provided for fine registration, and the accuracy and the efficiency of the fine registration are improved;

4. the two-step fine registration is mainly realized by a normal distribution transformation algorithm (NDT) and a closest point iteration algorithm (ICP) in an open source Point Cloud Library (PCL), firstly, the fine registration algorithm of the normal distribution transformation algorithm (NDT) is used for carrying out point cloud registration of two adjacent frames, then the closest point iteration algorithm (ICP) is used for correcting the position and posture of the point cloud of the two frames, and point cloud data obtained by multiple times of measurement are unified to the same coordinate system.

The point cloud mapping is implemented by splicing mass point cloud data acquired by a laser radar through a processor to form a visual image, and the visual image is displayed in the form of a top outline diagram, a left outline diagram and a right outline diagram of the rolling stock, and the process is as follows:

as shown in fig. 9, the OpenGL software is adopted to respectively perform the splicing processing on the point cloud data at the top of the rolling stock, the point cloud data at the left side and the point cloud data at the right side, and after the splicing is completed, the three position images are combined into one image to be displayed;

as shown in fig. 10, in the determination of the overrun portion, the three-dimensional point cloud data is divided into a left part, a right part and a top part 3, then each point of each part is determined from high to low in overrun level, only the highest overrun level of the point is reserved, if the number of overrun points of a certain part exceeds a certain scale, the level overrun is determined to occur, only the highest overrun level of the part is reserved for recording, in the three-dimensional point cloud image, the ground color is black, each point is displayed in different colors according to the overrun level, the normal point is displayed in white, the abnormal portion is displayed in gray, and the overrun portion display result;

as shown in fig. 11, the laser radar stores the point cloud mapping early warning information of the collected train outline data into the database, and the client software integrates the train type, the train number, the train speed, the outline point cloud image and other related information, so as to provide convenient and rapid train detection service for train detectors.

What needs to be supplemented is: setting of a geometric overrun judgment standard: according to the precondition specified by the railway transfinite overweight transport rule in China, after goods are loaded, the system measures the vehicles on a horizontal line and in a stopping state, so that the transfinite level is adjusted as follows: the original second-level overrun is specified as first-level overrun; the original secondary limit and the building limit are specified as super overrun; greater than the building clearance and prescribe as the building transfinite, above clearance standard data are saved in the database, call into the memory after the procedure starts, supply to transfinite and judge the use, for the convenience personnel fix a position and look over transfinite position, judge respectively after this system divides whole section vehicle three-dimensional point cloud data into the triplex, the triplex is respectively: vehicle left side, vehicle right side, and vehicle roof.

Specifically, the work flow of the vehicle inspection application platform which is a geometric overrun detection application software platform is to automatically acquire information such as train passing times, train numbers, train types, train speeds, passing time and the like; acquiring and processing the data of the outline dimension of the railway wagon, completing data segmentation by taking a carriage as a unit, and automatically generating a left side view, a right side view and a top view of the outline of the train; and the inspection personnel realize the vehicle inspection operation through client software.

It should be noted that: the system database of the data storage server is provided with three database relation tables including a passing car basic information table, a car checking information table and a user information table, the car basic information table is in TCP long connection with a front-end railway car number identification and detection system and a passing car detection and car division system in a Socket communication mode, and after the connection is established, basic information of each passing car is analyzed from received data and written into the database when a train passes through.

The computer device for client software setting of the geometry overrun detection application software platform comprises a processor and a memory, wherein the memory is coupled with the processor and used for storing the client software; the system comprises a user login module, a vehicle detection list module, an overrun state vehicle detection main interface, a report module and a user setting module, wherein the user login module, the vehicle detection list module, the overrun state vehicle detection main interface, the report module and the user setting module are interacted with computer equipment.

In this embodiment, the user login module is used as an entry module of the railway freight car loading state monitoring system, and the user can successfully log in after inputting a correct user name and a correct password. After the user inputs an account and a password, the system sends the information input by the user to the background server, the verification module verifies the information input by the user, the verification also relates to a verification code of the authority, and only the account party passing the verification can enter the system;

the vehicle detection list module is used for entering a vehicle detection list interface after logging in the system, mainly displaying the mounting position corresponding to front-end video and data acquisition, recording and listing information such as the number of vehicles, the type of vehicles, the number of vehicles, the vehicle detection state, the passing time of the vehicles and the like of each carriage in detail in the list part, displaying the latest vehicle passing data in the reverse order in the list, and performing vehicle detection operation on the carriage by clicking a right vehicle detection button by a vehicle detector;

the report module generates a vehicle-out quality report according to dates according to the daily report standard of railway freight safety inspection, and exports the report into an EXCEL file format, so that a vehicle inspector can count the daily vehicle inspection conditions conveniently, and the working efficiency is improved;

the user setting module is mainly used for adding and deleting system users by a system administrator, and can also modify user information such as user login passwords and the like;

the main interface of the overrun inspection mainly provides a quick and convenient inspection service for a goods inspector, the main interface is divided into three contour point cloud images of the left side, the top and the right side of each carriage, the three contour point cloud images are displayed in a simulated vehicle-passing rolling mode, the upper part of the right side is inquired according to the date and the number of vehicles, when a certain date is selected, all the number of vehicles in the day are displayed in a number list below, if the goods inspector needs to check the detailed overrun condition of each carriage in a certain number of vehicles, the number of vehicles is double-clicked by a mouse, the information of each carriage of the train is displayed in the lower vehicle list, the vehicles with abnormal conditions are displayed in red, early warning is carried out on the vehicle inspector, and the vehicle inspector can quickly select a single carriage needing to be checked in the vehicle number list.

The working principle and the using process of the invention are as follows: the dynamic scanning of the traveling train by the laser radar can be seen through field application and verification, the size data of the outline of the train can be scanned, the data are processed to generate point cloud images of 3 parts of the left side, the top and the right side of a single carriage, the point cloud images can show the details of the outline of the appearance of the train body, the common abnormal conditions such as the left side overrun, the top overrun, the right side overrun and the unclosed train door of the loading state of the train can be automatically judged according to the railway overrun overweight cargo transportation rule of a railway general company as the overrun judgment standard, and the abnormal parts are positioned. The car inspection personnel can conveniently and quickly realize the over-limit car inspection operation by using the car inspection application platform software, thereby effectively reducing the labor intensity, improving the working efficiency and improving the quality of goods inspection.

It is to be understood that the above-described embodiments of the present invention are merely illustrative of or explaining the principles of the invention and are not to be construed as limiting the invention. Therefore, any modification, equivalent replacement, improvement and the like made without departing from the spirit and scope of the present invention should be included in the protection scope of the present invention. Further, it is intended that the appended claims cover all such variations and modifications as fall within the scope and boundaries of the appended claims or the equivalents of such scope and boundaries.