CN111414503A

CN111414503A - Wood accumulation detection system based on block chain and CIM

Info

Publication number: CN111414503A
Application number: CN202010383906.1A
Authority: CN
Inventors: 刘克建; 何全芳
Original assignee: Individual
Current assignee: Individual
Priority date: 2020-05-08
Filing date: 2020-05-08
Publication date: 2020-07-14

Abstract

The invention discloses a wood accumulation detection system based on a block chain and CIM. The system comprises a wood region segmentation encoder, a wood region segmentation decoder, a noise suppression unit, a wood accumulation height estimation encoder, a full connecting layer, a wood region area estimation unit and a comprehensive analysis unit, and further comprises a calculation cluster, wherein all nodes in the calculation cluster load parameters required by a wood accumulation detection depth neural network, for image data sent by an image acquisition unit, wood accumulation detection block chain private chains are configured in the calculation cluster, network reasoning is executed, a wood accumulation comprehensive risk estimation result is obtained, and the wood accumulation comprehensive risk estimation result is sent to an urban construction site region information model. By utilizing the method and the device, in the prediction of the inflammable risk of the wood, not only result feedback is diversified, but also the practicability of the detection result, the precision of the detection result and the safety performance in the data processing process are improved.

Description

Wood accumulation detection system based on block chain and CIM

Technical Field

The invention relates to the technical field of block chains, artificial intelligence, CIM and intelligent construction sites, in particular to a wood accumulation detection system based on the block chains and the CIM.

Background

Wood is an indispensable important material in modern construction sites, and wood used in the construction sites at present is treated by various procedures such as antiseptic treatment, finished product assembly, surface modification and the like. Whether wood itself or wood shavings, polishing dust and flammable and combustible liquid glue materials are present in the wood, polishing dust and flammable and combustible liquid glue materials, a great fire hazard is generated. Once a fire breaks out, due to the nature of the material, the fire burns violently and spreads quickly, with unforeseeable consequences.

At present, monitoring cameras are arranged in a construction site, but the cameras only record and do not judge, the functions of the cameras are not fully exerted, abnormal conditions can be investigated and evidence can be obtained only through subsequent video playback, collected images are not processed, and real-time judgment and alarm cannot be achieved.

Some schemes use image information of a camera to detect the stacking condition of the wood, but most of the schemes adopt fixed threshold segmentation according to the color of the wood. Due to the influence of ambient light, the stability and accuracy of the method need to be improved. Moreover, these solutions often do not allow an assessment of the magnitude of the wood pile.

In recent years, deep learning methods are also adopted to classify images so as to determine whether wood accumulation exists in a current picture. If only judging whether wood is piled up or not, the estimation is often not accurate enough, and is not beneficial to large-scale practical application. Moreover, the current supervision system usually only returns an abnormal result, and the result feedback is single. In addition, the hardware system used for calculation is easy to leak information and has low safety performance. Therefore, the existing wood accumulation detection field has the problems of low detection precision, poor detection result practicability, single result feedback and lower safety performance.

Disclosure of Invention

The invention aims to provide a wood accumulation detection system based on a block chain and CIM (common information model), aiming at the defects in the prior art, not only result feedback is diversified, but also detection precision, detection result practicability and safety performance in a data processing process are improved.

A block chain and CIM based wood accumulation detection system, the system comprising:

the wood region segmentation unit is used for analyzing the color image of the monitored region to obtain a wood region mask and comprises a wood region segmentation encoder and a wood region segmentation decoder;

the noise suppression unit is used for performing median filtering on a feature map obtained by multiplying the monitored area depth map and the wood area mask point to obtain a wood area depth map;

the wood stacking height estimation encoder is used for extracting the characteristics of the depth map of the wood area to obtain a wood stacking height characteristic map;

the full-connection layer is used for carrying out weighted classification on the wood stacking height characteristic diagram to obtain a wood stacking height estimation result;

the wood region area estimation unit is used for carrying out area analysis on the wood region mask to obtain a wood region area estimation result;

the comprehensive analysis unit is used for integrating the wood stacking height estimation result and the wood area estimation result to obtain a wood stacking comprehensive risk estimation result;

the system further comprises a computing cluster, for the image data sent by the image acquisition unit, a wood accumulation detection block chain private chain is configured in the computing cluster, network reasoning is executed, a wood accumulation comprehensive risk estimation result is obtained, and the wood accumulation comprehensive risk estimation result is sent to the urban construction site area information model.

Further, the wood region segmentation encoder is used for encoding and extracting features of the color image in the monitored region to obtain a wood region semantic feature map;

the wood region segmentation decoder is used for decoding and representing the semantic feature map of the wood region to obtain a wood region mask.

Further, for the image data sent by the image acquisition unit, configuring a wood accumulation detection block chain private chain in the computing cluster comprises:

the wood region segmentation encoder, the wood region segmentation decoder, the noise suppression unit, the wood stacking height estimation encoder, the full connecting layer, the wood region area estimation unit and the comprehensive analysis unit are main component modules of the wood stacking detection depth neural network;

calculating parameters required by loading the wood accumulation detection deep neural network on all nodes in the cluster;

for image data sent by an image acquisition unit, selecting a plurality of available nodes from a computing cluster, taking parameters required by a wood accumulation detection depth neural network, which are respectively distributed in different available nodes, as block data of corresponding nodes by a wood area segmentation encoder, a wood area segmentation decoder, a noise suppression unit, a wood accumulation height estimation encoder, a full connection layer, a wood area estimation unit and a comprehensive analysis unit, and connecting blocks in the nodes according to a wood accumulation detection depth neural network reasoning sequence to generate a wood accumulation detection block chain private chain.

Further, the block in the private chain of block chains encrypts the neural network inference intermediate result data to be transmitted to the next block, and decrypts the neural network inference intermediate result data received from the previous block.

Furthermore, the block adopts a tensor confusion encryption mechanism to perform encryption and decryption operations.

Further, the tensor obfuscated encryption mechanism includes:

randomly generating C groups of random numbers, wherein each group comprises two random numerical values, and C is the channel number of the output tensor of the node where the block is located;

and performing cyclic shift operation on the width direction and the height direction of the tensor according to the two random numerical values.

Further, a wood region segmentation encoder, a wood region segmentation decoder, a noise suppression unit, a wood stacking height estimation encoder, a full connection layer, a wood region area estimation unit and a comprehensive analysis unit are respectively and properly subdivided, parameters of each subdivided module distributed in different nodes are used as block data, and a wood stacking detection block chain private chain is generated according to a wood stacking detection depth neural network reasoning sequence.

Further, the image data acquired by the image acquisition unit comprises a color image and a depth image of the monitored area.

Further, the system further comprises a visualization unit, the urban construction site region information model comprises construction site scene modeling information, monitoring region information and a wood accumulation risk estimation result in the urban region, and the visualization unit is used for visualizing the urban construction site region information model.

Further, the visualization unit comprises:

the initialization module is used for acquiring the modeling information of each construction site scene from the urban construction site region information model, rendering the urban construction site region information model by combining the WebGIS technology, and displaying the rendered urban construction site region information model on a foreground Web page to obtain an initial display result of the urban construction site region information model;

the data acquisition module is used for acquiring monitoring area information and a wood accumulation risk estimation result from the urban construction site area information model;

and the wood accumulation visualization module is used for matching the monitoring area information and the wood accumulation risk estimation result to the initial display result according to the geographical position of the monitoring area information, and carrying out warning marking on the monitoring area according to the abnormal early warning information.

Compared with the prior art, the invention has the following beneficial effects:

1. the method adopts the deep neural network to analyze the wood stacking height and the wood area, and compared with the traditional threshold segmentation detection method for wood stacking, the method uses a large number of samples, has better generalization performance, and improves the stability and accuracy of the system.

2. According to the wood accumulation detection method, the wood accumulation area and the wood accumulation height are considered, the comprehensive wood accumulation risk estimation result is obtained from two aspects, the risk caused by the wood accumulation height can be reflected, the probability of hidden dangers such as fire disasters possibly caused by the wood accumulation area can be reflected to a certain extent, and therefore the detection result is higher in practicability.

3. The method is based on the block chain technology, reasonably divides the deep neural network for wood stacking detection, dynamically generates the block chain private chain aiming at each network inference request, and compared with the traditional single-machine execution, not only improves the parallel performance of the system, but also has better fault-tolerant performance.

4. The block chain private chain is generated in real time according to available nodes in the computing cluster, is not easy to be tampered and attacked, and has higher confidentiality.

5. The invention encrypts the data between the private chains of the network inference block chain, prevents the leakage of the transmission data between the private chain blocks of the block chain, ensures the confidentiality of the transmission data, has convenient tensor confusion encryption operation and small calculation amount, and can not increase the system burden while improving the confidentiality.

6. The method is based on CIM technology to design the city construction site area information model to store the wood accumulation detection result and visualize the city construction site area information model, compared with the traditional result feedback, the feedback result of the method is more diversified, and comprises construction site area three-dimensional display, warning marks and monitoring area images, so that the supervisor can more clearly and definitely know the risk estimation condition caused by wood accumulation in the construction site.

Drawings

FIG. 1 is a diagram of a neural network architecture of the system of the present invention.

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 below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

The invention provides a wood accumulation detection system based on a block chain and CIM. Firstly, image features are extracted through a color area-array camera, then the image features are divided into two branches, one branch is responsible for a mask image of wood accumulation, the other branch is responsible for obtaining height information, the accumulation height information and the area size information of the accumulation are fused, and finally the information of the wood accumulation risk degree is obtained. FIG. 1 is a diagram of a neural network architecture of the system of the present invention. The following description will be made by way of specific examples.

The first embodiment is as follows:

a wood accumulation detection system based on a block chain and a CIM is used for performing wood accumulation detection based on an urban construction site area information model. The urban construction site area information model comprises construction site scene modeling information, monitoring area information and wood accumulation detection data.

Specifically, a Building Information model (Building Information Modeling) is a new datamation tool applied to design, construction and management of Building engineering, and integrates all Building data in a Building project cycle. The visualization of the building information model can be realized by using visualization software in combination with technologies such as WebGIS and the like, and convenience is provided for building design and building management. The CIM (City Information Modeling, City Information model) is a further upgrade of the BIM, extends the Modeling range from a single building model or a plurality of building models to the three-dimensional Modeling of the whole City, and can perform three-dimensional Modeling, displaying and managing on the scenes of buildings, traffic, roads and the like of the whole City.

The invention combines CIM to display the wood accumulation detection result in real time and provides early warning information for the monitoring personnel. Therefore, the invention designs an urban construction site area information model. The city construction site region information model is based on CIM technology and comprises construction site scene modeling information, monitoring region information and wood accumulation risk estimation results in a city or a certain region of the city. The building site scene modeling information comprises building site building information, building site building material placement information, worker work area information and other various building site scene information, the type information simultaneously contains corresponding geographic position information, and the three-dimensional scene of the building site can be restored and displayed through visual software by combining a Web GIS technology. The monitoring area information comprises monitoring area geographical position information, monitoring visual sensor geographical position information, images shot by the monitoring visual sensors and a coordinate transformation matrix. The monitored area information is used to restore the monitored area image in the visualized building information model of the worksite area. The coordinate change matrix is used for converting the monitoring image into an image with better visualization effect, such as a top view and the like. And integrating the wood accumulation risk estimation result into a matched monitoring area for the supervision personnel to check. In the invention, the wood accumulation detection result is transmitted to the urban construction site area information model in real time. The wood accumulation risk estimation result comprises a wood accumulation area estimation result, a height estimation result, a comprehensive risk estimation result and the like.

The following is a detailed description of how to obtain a comprehensive risk assessment of wood accumulation. According to the method, a series of data collected by the sensor are processed, and the characteristics of the sensing information of the sensor are extracted by adopting a deep learning method, so that the purpose of estimating the comprehensive risk of wood accumulation is achieved. The wood stacking detection deep neural network comprises a wood region semantic segmentation sub-network and a wood stacking height estimation sub-network. The wood region semantic segmentation sub-network comprises a wood region segmentation encoder and a wood region segmentation decoder; the wood region segmentation encoder is used for encoding and extracting features of the color image in the monitored region to obtain a wood region semantic feature map; and the wood region segmentation decoder is used for decoding and representing the semantic feature map of the wood region to obtain a wood region mask. The wood stacking height estimation sub-network comprises a wood stacking height encoder and a full connection layer; the wood stacking height estimation encoder is used for extracting the characteristics of the depth map of the wood area to obtain a wood stacking height characteristic map; and the full connecting layer is used for carrying out weighted classification on the wood stacking height characteristic diagram to obtain a wood stacking height estimation result. The wood area depth map is obtained through a noise suppression unit, specifically, the noise suppression unit is used for performing median filtering on a feature map obtained by point-to-point multiplying the monitored area depth map and the wood area mask to obtain the wood area depth map.

It should be noted that, only if the presence or absence of wood accumulation is determined, the fire hazard can be estimated only from a single aspect, and the estimation is often not accurate enough, which is not beneficial to large-scale practical application. And the grade information is piled up to timber and is amalgamated with factors such as the area that timber was piled up, pile up height, and the real dangerous grade that has reflected the accumulational fire hazard that causes of timber, the overall planning and the arrangement of building site construction are done well to the help administrator that this grade information can be better, improve the efficiency of construction.

In the implementation, the RGB-D camera is adopted to collect the images of the monitoring area, and the RGBD information can be firstly divided into an RGB three-channel image and a depth image of the monitoring area through channel split operation.

The invention relates to a wood region semantic segmentation subnetwork, which is shown in figure 1: firstly, a wood region semantic feature map FeatureMapa of a wood mask is extracted through a wood region segmentation encoder EncodeA, the obtained wood semantic feature map is input into a wood region segmentation decoder DecodeA, and the wood region mask can be obtained. It should be noted that the mask of the wood area not only contains whether the current pixel is wood, but also indirectly obtains the information of the specific geographic position coordinate of the wood stacking.

In the training process, the input of the wood region semantic segmentation sub-network is a color picture, the output is the mask information of the wood stacking region, and the monitoring information is a binary image which is labeled manually. The binary image indicates that the pixel value of the wood region is marked as 1, and other background pixel information is marked as 0. And training the network by adopting cross entropy loss to obtain the shade information of the wood. And finishing the extraction of the wood region mask information.

The wood height estimation subnetwork of the invention, as shown in fig. 1: first, multiplication of the wood region mask and the pixel value of each point of the depth image is required. For example, the following steps are carried out: the mask information is a binary image composed of 0 and 1, where 0 represents background information and 1 represents wood mask information. For each pixel location, the mask information is multiplied by the thermal imaging information, any number multiplied by 0 remains 0, and any number multiplied by 1 is the numerical value itself. Finally, the representation information of the wood region is enhanced in the depth map, and the representation information of the background region is suppressed. Thereby obtaining a depth map of the wood region.

Further, since the depth values at some points in the depth image are obviously different from the depth values of the surrounding points, which are noise points in the depth image, if the unprocessed input network learns, the wrong depth features of the wood region can be extracted. Therefore, the invention adopts a median filtering method to filter noise points and protect edge information, thereby extracting accurate depth characteristic information of the wood accumulation area.

Further, the characteristics of the depth map of the wood region are extracted through a wood stacking height estimation encoder EncoderB, and then a current wood stacking height estimation value is finally obtained through a full connection layer FC.

In the training process, the input of the wood stacking height estimation sub-network is a depth map of the current area, the output is a wood stacking height estimation value, and the supervision information is the wood flammability risk of each image sample which is labeled manually. And updating parameters by adopting common cross entropy loss, classifying the input depth map, and finally obtaining a wood stacking height estimation result. It should be noted that the present invention divides the wood stacking height into four levels: index 0 represents no wood stacking, index 1 represents low stacking height, index 2 represents medium stacking height, and index 3 represents higher stacking height. Specifically, the grade division standard can be set by an implementer according to implementation requirements. For example, a height of less than 1 meter is a low stacking height, a height of between 1 and 3 meters is a medium stacking height, and a height of more than 3 meters is a high stacking height.

It should be noted that the consideration of the comprehensive risk of wood stacking includes not only the stacking height, but also the information of the area size of the current wood stacking. The higher the pile-up, the higher the risk of collapse and drop; correspondingly, under the premise of a certain stacking height, the larger the stacking area is, the higher the fire risk level is.

Therefore, the wood region area estimation unit is arranged, and after the binary image of the wood region mask is obtained, the area information of the wood stacking plane is obtained through a contour searching and contour area calculation post-processing mode. Further, an area grade standard may be set according to a ratio between the area of the wood region and the area of the camera view, and the area grade information of the wood stacking may be determined according to the estimation result of the wood stacking area.

In order to integrate the analysis results of the wood stacking height and the wood area, the comprehensive analysis unit is arranged, and the wood stacking height estimation result and the wood area estimation result can be integrated to obtain the wood stacking comprehensive risk estimation result. A simple comprehensive way is to add the height grade of the wood pile and the area grade of the wood pile through add operation to obtain the final wood pile grade, i.e. the wood pile comprehensive risk estimation result. In addition, the implementer can also set corresponding weights for the two risk factors according to the proportion of the two risk factors in the risk assessment.

And obtaining the comprehensive risk estimation result of the wood accumulation.

In the neural network, an encoder is used for carrying out feature extraction on input multi-channel two-dimensional data. The encoder may employ CNN Block, Res Block, or the like. The encoder and the decoder can be realized in various ways, in order to take the size of a large target into consideration, the invention proposes to adopt an hourglass network to extract features, and an implementer can also select a proper module design in a neural network according to the size of an image and the occupation of a video memory, such as Residual Block, Bottleneck Block, CNN Block and the like. The encoder, the decoder and the full-connection network of the invention adopt which network design, an implementer can select according to the specific implementation requirements, and the modularization idea is the protection content of the invention.

In order to improve the confidentiality of the system and prevent data leakage, the invention designs the private chain of the block chain by combining the block chain technology.

The present invention is described in detail herein in connection with the block chain technique and the DNN technique. The block chain adopts block division data, a chain data structure is used, the data are used as blocks for verification and storage, the whole data structure is summarized, centralized hardware and management mechanisms do not exist, and decentralization is achieved. The block chain technology of the first generation is mainly applied to a distributed account book, the block chain technology of the second generation mainly realizes an intelligent contract, and the block chain idea and other field technologies can be combined by the block chain technology of the third generation, so that more and more presentation forms exist, and more emphasis is placed on system function service. The block chain private chain completely inherits the characteristics of the public chain, is not bound by a game mechanism, focuses more on data transmission and encryption and decryption processing of practical application, and can be better combined with technologies in other fields. For deep neural network computation in artificial intelligence, it is not necessary to store intermediate result data, and chained logic is retained to match the principle of neural network forward propagation.

The invention considers the problem that the contents of the image data in the plaintext are leaked in the uploading process and the processing process when the data are directly uploaded to be processed, so that the invention uses the form of a block chain private chain, takes different modules of a deep neural network as blocks, carries out dispersed reasoning and carries out encryption processing on the data transmitted among the blocks, thereby realizing excellent performances of parallel reasoning, fault tolerance and data leakage prevention.

Based on the thought, the wood accumulation detection depth neural network is divided into modules, and a wood region division encoder, a wood region division decoder, a noise suppression unit, a wood accumulation height estimation encoder, a full connection layer, a wood region area estimation unit and a comprehensive analysis unit are used as different modules of the network. Thus, a wood accumulation detection deep neural network inference chain can be obtained according to the inference sequence of the neural network shown in fig. 1. It should be noted that, strictly speaking, the noise suppression unit is an image processing method of the network on the intermediate processing result, and the wood region area estimation unit and the comprehensive analysis unit belong to a post-processing mode on the network output, but in a broad sense, three units can also be attributed to the wood accumulation detection deep neural network. In order to make the relationship between each unit and the subsequent block chain private chain clearer, the three units are classified into the deep neural network in a generalized exposition mode.

The system also comprises a calculation cluster, wherein all nodes in the calculation cluster are loaded with the weight and the parameters required by the deep neural network for wood accumulation detection; and selecting a plurality of available nodes from the computing cluster according to each wood accumulation detection depth neural network reasoning request, and taking the wood region segmentation encoder, the wood region segmentation decoder, the noise suppression unit, the wood accumulation height estimation encoder, the full connection layer, the wood region area estimation unit and the weight and the parameters required by the comprehensive analysis unit which are respectively distributed in different available nodes as block data of the corresponding nodes. Thus, a wood region segmentation encoder block, a wood region segmentation decoder block, a noise suppression unit block, a wood stacking height estimation encoder block, a full link layer block, a wood region area estimation unit block and a comprehensive analysis unit block distributed at different nodes can be obtained. And connecting the blocks according to the wood accumulation detection depth neural network reasoning sequence to generate a wood accumulation detection block chain private chain, and executing the wood accumulation detection depth neural network reasoning. The network link order is consistent with the block chain private link order, as shown in fig. 1.

When selecting available nodes and performing node sorting, preferably, the available nodes in the computing cluster are randomly sorted, and computing nodes with the same number as the number of blocks are selected from the sorted nodes. Specifically, a random number is generated for each computing device, N (N is the number of computing nodes), and the N random numbers are generated into their ranking indexes in the order from small to large, the foregoing step has divided the network into 7 modules, so that the ranking index of the top 7 random numbers in the random number sequence is selected, and the computing devices corresponding to the indexes constitute the nodes selected by chain inference. For example, the generated random numbers are [11,7,2,3,6,8,10,1,4,5], respectively, and [10,7,2,3,6,8,9,1,4,5] is obtained by sorting in descending order, and then the nodes corresponding to the [10,7,2,3,6,8,9] indexes are sequentially selected. Selecting a first node from the selected available nodes, and taking parameters such as weights required by the timber region segmentation encoders in the nodes as block data; and taking a second node, taking parameters such as the weight required by the wood region segmentation decoder in the node as block data, linking the block data with the previous block, and by analogy, generating a corresponding wood accumulation detection block chain private chain according to a neural network reasoning sequence. Therefore, a plurality of wood accumulation detection block private chains generated aiming at different requests can exist in the computing cluster at the same time, and the block private chain is dynamically generated, is not easy to be cracked by attacks, and has better confidentiality. And after the block chain private chain is obtained according to the inference sequence, performing network inference calculation on the inference request according to the inference sequence.

Some modules of the neural network are difficult to be put into one node at one time, namely, the computation of some modules is large, and the computation is difficult to be completed in a short time. Therefore, the blocks, i.e., the above-mentioned encoder, decoder, etc., units can be further divided in advance. The granularity of the further segmentation can be adjusted by the implementer according to the specific implementation. In the present embodiment, it is preferable that the wood region division encoder is divided into 3 blocks, the wood region division decoder is divided into 3 blocks, the wood stacking height estimation encoder is divided into 3 blocks, the full-link layer is divided into 4 blocks, and the remaining units may not be divided. Therefore, according to the inference sequence of the neural network, a more subdivided wood volume detection deep neural network inference chain can be obtained. Correspondingly, aiming at each wood accumulation detection deep neural network reasoning request, selecting a plurality of available nodes from a computing cluster, taking weights and parameters required by sub-modules which are respectively distributed in different available nodes after subdivision as block data of corresponding nodes, and generating a wood accumulation detection block chain private chain according to a wood accumulation detection deep neural network reasoning sequence.

Meanwhile, the image acquisition unit can be used as a block and added into the block chain private chain, that is, the RGBD camera end is added into the block chain private chain, and the camera parameters therein are used as block data. The purpose of doing so is to guarantee that the output of image acquisition unit is secret through subsequent encryption strategy, is difficult for intercepting, falsifying. The node where the monitoring center receiving the wood accumulation comprehensive risk estimation result is located can also be added into the block chain private chain, so that the purpose of ensuring the safety of data transmission between the monitoring center and the computing cluster is achieved.

Further, in order to ensure the integrity and security of data received by each block and prevent data from being attacked and tampered during transmission, data transmitted between blocks needs to be encrypted. That is, data transmission from private link node to node of the block chain requires the use of encryption, preferably using a tensor obfuscated encryption mechanism.

Specifically, the encryption is completed by carrying out tensor confusion on the feature maps output by the current block or the computing node, wherein the specific confusion mode is that for each channel of the current feature map, the size of the channel is H × W, a group of random number combinations are selected to carry out cyclic shift on the current feature map, each group has two random numerical values, C groups are shared, the operation is respectively carried out on the width and the height of the feature map, and C represents the number of the channels of the current feature map.

For example, assuming that a shift operation is currently performed on the feature map with index i, the current random number combination is (m, n), where m < W and n < H. And when m is an odd number, circularly shifting to the left by the shift step length of m, otherwise, circularly shifting to the right. Assuming that the data in the horizontal direction of the original image is 011010001, when the value of m is 2 and even, the step length of right cyclic shift is 2 according to the rule, and after the shift, the step length is 010110100; when m is 3, odd, left shifted 110100010. The image vertical direction is the same as that: and when n is an odd number, performing downward cyclic shift, otherwise performing upward cyclic shift, wherein the shift step size is n. The intermediate result is one-dimensional, and cyclic shift in the left-right direction may be performed using the parameter m. When the intermediate result is single data, a universal encryption algorithm is adopted.

And applying the encryption algorithm to the data to be transmitted among all the blocks to finish the encryption operation of the data. After the data is input into the next block, the data is decrypted according to the reverse reasoning of the encryption rule, so that the encrypted transmission of the data between the computing nodes is completed.

It should be noted that the encryption parameters required for transmitting data between the wood stacking detection deep neural network modules, i.e. between the blocks, should be generated by the trusted nodes. The trusted node may be fixed, for example, a master node may be configured to broadcast encryption parameters and allocation tables periodically to prevent the encryption parameters from being broken. However, the fixed trusted node is vulnerable to attack, and therefore, preferably, the encryption parameter required by the next inference request is generated by the node where the last block of the private chain of the block chain is executed in the computing cluster, and is broadcast to all nodes in the computing cluster, the terminal cluster and the monitoring center.

In order to visually present the CIM information state of a construction site area and enable a construction site manager to visually acquire the camera perception information of the current area and the information of whether wood is accumulated or not through network prediction, the invention combines a Web GIS visualization technology to display the CIM information model on Web through the Web GIS technology.

And the visualization unit is used for acquiring data from the urban construction site region information model, rendering the urban construction site region information model by combining the Web GIS, and displaying construction site wood accumulation conditions on a foreground Web page. The visualization unit includes: and the initialization module is used for acquiring the building site scene modeling information from the urban building site region information model, rendering the urban building site region information model by combining the Web GIS technology, and displaying on a foreground Web page to obtain an initial display result of the urban building site region information model. And the data acquisition module is used for acquiring monitoring area information and a wood accumulation risk estimation result from the urban construction site area information model. And the wood accumulation risk visualization module is used for matching the monitoring area information and the wood accumulation risk estimation result to the initial display result according to the geographical position of the monitoring area information, and carrying out warning marking on the monitoring area according to the abnormal early warning information.

When the risk of wood accumulation is predicted to be high, alarm information can be sent out, and the alarm information is marked in the CIM visualization result according to the position of the monitored area, so that the supervision personnel can take corresponding preventive and emergency measures according to the geographic position coordinates of the supervision personnel.

The above embodiments are merely preferred embodiments of the present invention, which should not be construed as limiting the present invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims

1. A timber pile-up detecting system based on block chain and CIM is characterized in that the system comprises:

2. The system according to claim 1, wherein the wood region segmentation encoder is configured to encode and extract features from the color image of the monitored region to obtain a semantic feature map of the wood region;

3. The system of claim 1, wherein configuring in the computing cluster the wood accumulation detection block chain private chain for the image data sent by the image acquisition unit comprises:

for image data sent by an image acquisition unit, selecting a plurality of available nodes from a computing cluster, taking parameters required by a wood accumulation detection depth neural network, which are respectively distributed in different available nodes, as block data of corresponding nodes by a wood area division encoder, a wood area division decoder, a noise removal unit, a wood accumulation height estimation encoder, a full connection layer, a wood area estimation unit and a comprehensive analysis unit, and connecting blocks in the nodes according to a wood accumulation detection depth neural network reasoning sequence to generate a wood accumulation detection block chain private chain.

4. The system of any one of claims 1-3, wherein a tile in the private chain of tiles encrypts neural network inference intermediate result data that it is to transmit to a next tile, and decrypts neural network inference intermediate result data that it receives from a previous tile.

5. The system of claim 4, wherein the block uses tensor obfuscated encryption to perform encryption and decryption operations.

6. The system of claim 5, wherein the tensor obfuscated encryption mechanism comprises:

randomly generating C groups of random numbers, wherein each group comprises two random numerical values;

7. The system according to claim 1, wherein the wood region segmentation encoder, the wood region segmentation decoder, the noise suppression unit, the wood accumulation height estimation encoder, the full link layer, the wood region area estimation unit, and the comprehensive analysis unit are respectively and appropriately subdivided, and parameters of each subdivided module respectively distributed at different nodes are used as block data to generate the wood accumulation detection block chain private chain according to the wood accumulation detection depth neural network inference sequence.

8. The system of claim 1, wherein the image data acquired by the image acquisition unit comprises a monitored area color image and a monitored area depth image.

9. The system of claim 1, wherein the system further comprises a visualization unit, the urban worksite area information model comprises site scene modeling information, monitoring area information, and a wood accumulation comprehensive risk estimation result in the urban region, and the urban worksite area information model is visualized by the visualization unit.

10. The system of claim 9, wherein the visualization unit comprises:

the initialization module is used for acquiring the modeling information of each construction site scene from the urban construction site region information model, rendering the urban construction site region information model by combining the Web GIS technology, and displaying the rendered urban construction site region information model on a foreground Web page to obtain an initial display result of the urban construction site region information model;