A multi-modal construction site no-entry area fence warning method and device

Abstract

The application discloses a kind of genetic algorithm-based multi-modal construction site forbidden area fence alarm method and device, the method includes inputting the plan of construction site, marking out construction site information;The plan is divided into several grid units, each grid unit is set as fence or blank;Target in fence area is identified, and genetic algorithm is used to extract target features to identify whether target enters fence area, and if so, an alarm prompt is sent out.The application adopts a multi-modal algorithm model based on genetic algorithm technology to implement a safety warning scheme for the fence in the forbidden area of the construction site, which can provide real-time, efficient and accurate safety warnings for personnel and vehicles inside the fence.

Description

Technical Field

[0001] This invention relates to the field of safety management technology for construction sites, specifically to a multimodal construction site restricted area fencing alarm method and device based on genetic algorithms. Background Technology

[0002] At construction sites, restricted areas must be established to ensure the safety of workers and the public. These areas are typically separated by fences. Traditional restricted area fencing technology is usually based on experience and manual design, failing to fully consider the characteristics and complexity of the construction site. Furthermore, due to the complexity of construction sites, traditional restricted area fencing technology often requires multiple adjustments and modifications, leading to extended construction periods, increased costs and risks, and increased safety risks. In electronic fence construction, the construction environment is often harsh, and the safety of on-site construction personnel cannot be guaranteed. In the event of an emergency, on-site construction personnel cannot be notified in a timely manner, resulting in many avoidable accidents. In this situation, anticipating the behavior of construction personnel and providing timely warnings are crucial.

[0003] Current research indicates that workers can monitor construction information data in real time through a monitoring system to understand the construction status. This work method requires the support of personnel positioning technology to track construction workers, transmit data to a computer, and thus determine their location. When alarm data is received, corresponding alarm measures are taken in a timely manner to ensure the safety of construction workers while monitoring the construction progress. However, at present, some common alarm algorithms can no longer meet the actual needs of electronic fence construction and have some shortcomings. Common virtual electronic fence technologies include pulse electronic fences, infrared electronic fences, intelligent pulse electronic fences, and tension electronic fences. Pulse electronic fences mainly consist of a pulse host and a front-end fence. The pulse host emits high-voltage low-frequency positive pulses and high-voltage negative pulses. The energy of the pulses, controlled within a certain energy range, will cause a strong stinging sensation to people who touch the fence, thus serving as a deterrent. However, this method may cause varying degrees of mental and physical harm to the human body. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a multimodal construction site restricted area fence alarm method based on genetic algorithm.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] In a first aspect, the present invention provides a multimodal construction site restricted area fencing alarm method based on a genetic algorithm, the method comprising:

[0007] Input a site plan and mark the site information;

[0008] Divide the floor plan into several grid cells, and set each grid cell as either a fence or a blank space;

[0009] The system identifies targets within the fenced area and uses a genetic algorithm to extract target features to determine whether a target has entered the fenced area. If so, an alarm is issued.

[0010] Furthermore, the construction site information includes hazardous areas, entrances and exits, and pedestrian density.

[0011] Furthermore, target identification within the fenced area includes facial recognition, pedestrian recognition, and vehicle recognition.

[0012] Furthermore, the face recognition includes: extracting local facial features using a genetic algorithm, verifying the features using the eyes and forehead, and then using other local features as auxiliary tools for face verification.

[0013] Furthermore, the pedestrian identification includes: pedestrian feature extraction, and using a genetic algorithm to extract robust pedestrian features for identification.

[0014] Furthermore, the vehicle identification includes: extracting vehicle features from the construction site and using a genetic algorithm to extract robust vehicle features for identification.

[0015] Furthermore, the use of genetic algorithms to extract target features includes:

[0016] Initialize a population, where each individual represents a possible target feature;

[0017] Define a fitness function to evaluate the quality of each individual's genetic information characteristics;

[0018] Calculate the fitness value of each individual based on the fitness function, and sort them from high to low fitness values;

[0019] Based on the preset crossover and mutation probabilities, crossover and mutation operations are performed on individuals in the population to generate new individuals;

[0020] Two individuals are selected as parents based on their fitness values, and then a single crossover is performed at a random location to produce two offspring.

[0021] Perform a mutation operation on each bit of each offspring;

[0022] Add the newly generated N offspring to the population, and eliminate the N individuals with the lowest fitness values ​​to keep the population size unchanged; N is a positive integer;

[0023] Determine whether the preset termination condition has been met. If so, output the individual with the highest fitness value in the current population as the optimal feature.

[0024] Furthermore, the fitness function is:

[0025] f(x)=R_1\cdot I_1(x)+R_2\cdot I_2(x)+R_3\cdot I_3(x)-C_1\cdot L(x)-C_2\cdot S(x)

[0026] Where x represents a fence design; R_1, R_2, and R_3 represent the reward values ​​when the constraints are met; C_1 and C_2 represent the coefficients of the fence length and shape; I_1(x), I_2(x), and I_3(x) indicate whether the constraints are met, and if they are met, the value is 1.

[0027] Secondly, the present invention provides a multimodal construction site restricted area fence alarm device based on a genetic algorithm, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0028] Thirdly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the methods described above.

[0029] Compared with the prior art, the advantages of this invention are as follows:

[0030] This invention employs a multimodal algorithm model based on genetic algorithm technology to implement a safety early warning scheme for the fence of restricted areas in construction sites, which can provide real-time, efficient and accurate safety warnings for personnel and vehicles within the fence. Attached Figure Description

[0031] Figure 1 This is a flowchart of the multimodal construction site restricted area fence alarm method based on genetic algorithm provided in Embodiment 1 of the present invention;

[0032] Figure 2 This is a site plan of the construction area;

[0033] Figure 3 A flowchart for local feature recognition in face recognition;

[0034] Figure 4 Flowchart for vehicle identification algorithm;

[0035] Figure 5 Flowchart for alarm notifications at construction sites;

[0036] Figure 6 Here is an example of a crossover operation;

[0037] Figure 7 Here is an example of a mutation operation;

[0038] Figure 8 This is a schematic diagram of the composition of the multimodal construction site restricted area fence alarm device based on genetic algorithm provided in Embodiment 2 of the present invention. Detailed Implementation

[0039] Example:

[0040] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0041] See Figure 1 As shown in the figure, the multimodal construction site restricted area fence alarm method based on genetic algorithm provided in this embodiment mainly includes the following steps:

[0042] 101. Input a plan of the construction site and mark the construction site information, including hazardous areas, entrances and exits, and pedestrian density;

[0043] In one specific embodiment, the plan view of the construction site is as follows: Figure 2 As shown, the rectangular areas represent danger zones that need to be completely isolated; entrances and exits must remain unobstructed; other areas represent normal zones where free passage is permitted.

[0044] 102. Divide the plan into several grid cells, each of which can be set as a fence or blank;

[0045] 103. Identify targets within the fenced area, use a genetic algorithm to extract target features, and determine whether a target has entered the fenced area. If so, issue an alarm.

[0046] In this way, by identifying targets within the fenced area, collision warnings can be issued for targets within the fenced area, and timely alarms can be given to avoid construction safety accidents. This enables real-time, efficient, and accurate safety alarms for personnel and vehicles within the fenced area.

[0047] In one specific embodiment, target identification within the fenced area includes facial recognition, pedestrian recognition, and vehicle recognition, which can provide real-time, efficient, and accurate safety alerts for personnel and vehicles within the fenced area.

[0048] The facial recognition process is as follows Figure 3As shown, this involves using a genetic algorithm to extract local facial features, prioritizing the use of the eyes and forehead for verification, and then supplementing with other local features for facial verification. In this way, facial recognition is used to determine the entry and exit of personnel at entrances and exits, ensuring that there are no unauthorized personnel at the construction site.

[0049] Pedestrian recognition mainly includes pedestrian feature extraction, using genetic algorithms to extract robust pedestrian features for recognition.

[0050] The flowchart for pedestrian and vehicle recognition is as follows: Figure 4 As shown, this includes extracting vehicle features from the construction site and using a genetic algorithm to extract robust vehicle features for identification.

[0051] In this way, by identifying construction workers, vehicles, and equipment within the construction site, red danger zones are defined to prevent collisions with these zones. Furthermore, the system can calculate the speed of personnel, vehicles, and equipment and issue alarms for potential collisions. The alarm process is as follows: Figure 5 As shown

[0052] In one specific embodiment, step 103 above, using a genetic algorithm to extract target features includes:

[0053] Initialize a population, where each individual represents a potential target trait; define a fitness function to evaluate the quality of each individual's genetic information traits, considering the following factors:

[0054] 1) Feature representation discriminative power, which is the distance between features of different categories, the larger the better;

[0055] 2) Robustness of feature representation, that is, the generalizability of the feature representation; the stronger the generalizability, the better.

[0056] 3) The dimensionality of the features, i.e. the complexity of the feature representation, should be as simple as possible;

[0057] Considering the above three factors, the fitness function is cleverly designed as follows:

[0058] f(x)=R_1\cdot I_1(x)+R_2\cdot I_2(x)+R_3\cdot I_3(x)-C_1\cdot L(x)-C_2\cdot S(x)

[0059] Where x represents a fence scheme; R_1, R_2, and R_3 represent the reward values ​​when the constraints are met; C_1 and C_2 represent the coefficients of the fence length and shape; I_1(x), I_2(x), and I_3(x) indicate whether the constraints are met, with 1 if met and -1 otherwise; L(x) represents the total length of fence scheme x; and S(x) represents the number of connected components of fence scheme x.

[0060] Next, the fitness value of each individual is calculated according to the fitness function, and they are sorted from high to low fitness value. Assuming R_1=R_2=R_3=100 and C_1=C_2=10, Table 1 shows the fitness value and sorting results of each individual in the initial population.

[0061] Table 1. Fitness values ​​and ordination results for each individual in the initial population.

[0062] Serial Number fitness value Serial Number fitness value Serial Number fitness value Serial Number fitness value 1 0.210491364 26 0.815621707 51 0.478468888 76 0.902108961 2 0.10944786 27 0.035188114 52 0.709467359 77 0.059434928 3 0.638439841 28 0.966343993 53 0.272497267 78 0.409818199 4 0.853687362 29 0.595010781 54 0.877424369 79 0.587892834 5 0.570556214 30 0.358771883 55 0.360857276 80 0.142548179 6 0.502052088 81 0.862244777 56 0.364291104 81 0.715990235 7 0.189066614 82 0.582878054 57 0.994209936 82 0.689171812 8 0.15005482 83 0.177652006 58 0.802229817 83 0.943029907 9 0.499800281 34 0.970948447 59 0.829448786 84 0.218395782 10 0.693693628 35 0.571118382 60 0.626886293 85 0.549955475 11 0.281836682 36 0.369435957 61 0.046701361 86 0.145212599 12 0.622149652 37 0.114239417 62 0.019377535 87 0.583551492 13 0.899341425 38 0.179814371 63 0.587012089 88 0.482410267 14 0.724639818 39 0.2399873 64 0.010139686 89 0.10930135 15 0.710295938 40 0.154027526 65 0.811768526 90 0.893716679 16 0.113482429 41 0.904964684 66 0.696774497 91 0.180673751 17 0.724248136 42 0.580451393 67 0.08802112 92 0.668936088 18 0.889825144 43 0.725824767 68 0.502249295 93 0.700488372 19 0.102227846 44 0.35060997 69 0.909120725 94 0.375238426 20 0.926740977 45 0.067105422 70 0.854818053 95 0.368915201 21 0.342916975 46 0.952082838 71 0.211702279 96 0.628462995 22 0.826791122 47 0.659140207 72 0.224606349 97 0.224884482 23 0.183680503 48 0.172392983 73 0.01695914 98 0.249170362 24 0.812930736 49 0.072375592 74 0.683640276 99 0.037727581 25 0.364769827 50 0.070093404 75 0.024164064 100 0.099299176

[0063] Then, based on the preset crossover and mutation probabilities, crossover and mutation operations are performed on individuals in the population to generate new individuals. Assuming the crossover probability is 0.8 and the mutation probability is 0.1, the crossover and mutation process is as follows:

[0064] Crossover: Select two individuals as parents based on their fitness values, and then perform a single-point crossover at a random location to produce two offspring. Repeat this process until 10 offspring are produced. For example, if the 1st and 3rd individuals are selected as parents, and the 8th individual is randomly chosen as the crossover point, the resulting offspring will be as follows: Figure 6 As shown.

[0065] Thus, by adopting a single-point crossover approach, the encoding and computation process is simplified, computational complexity and memory consumption are reduced, and this approach is introduced into neural convolutional networks to optimize target features and improve robustness and distinguishability.

[0066] Mutation: A mutation operation is performed on each bit of each offspring, inverting the bit with a probability of 0.1. For example, if the first offspring is mutated, and the 2nd and 9th bits are randomly selected for mutation, the mutated offspring will look like this: Figure 7 As shown.

[0067] Next, the 10 newly generated offspring were added to the population, and the 10 individuals with the lowest fitness values ​​were culled, keeping the population size unchanged. Table 2 shows the fitness value and ranking results for each individual in the new population.

[0068] Table 2. Fitness values ​​and ordination results for each individual in the new population.

[0069]

[0070]

[0071] Determine if the preset termination conditions have been met, such as the maximum number of iterations, optimal solution stability, fitness threshold, etc. If so, output the individual with the highest fitness value in the current population as the optimal feature; otherwise, return to the previous step to continue iterating. Assuming the maximum number of iterations is 1000, after 1000 iterations, the optimal feature representation is obtained, satisfying all constraints, and the dimension of the feature representation is relatively good.

[0072] In summary, based on computer vision theory research, this invention uses genetic algorithms for pedestrian re-identification and face recognition, and provides a multimodal construction site restricted area fence alarm method based on genetic algorithms. This method can provide real-time, efficient and accurate alarms for the safety of personnel and vehicles within the fence, adapting to the safety needs of different sites.

[0073] Example 2:

[0074] See Figure 8 As shown, the multimodal construction site restricted area fence alarm device based on genetic algorithm provided in this embodiment includes a processor 81, a memory 82, and a computer program 83 stored in the memory 82 and executable on the processor 81, such as a multimodal construction site restricted area fence alarm program based on genetic algorithm. When the processor 81 executes the computer program 83, it implements the steps of Embodiment 1 described above, for example... Figure 1 The steps are shown.

[0075] For example, the computer program 83 can be divided into one or more modules / units, which are stored in the memory 82 and executed by the processor 81 to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program 83 in the multimodal construction site restricted area fence alarm device based on genetic algorithms.

[0076] The multimodal construction site restricted area fence alarm device based on genetic algorithms can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The multimodal construction site restricted area fence alarm device based on genetic algorithms may include, but is not limited to, a processor 81 and a memory 82. Those skilled in the art will understand that... Figure 5 This is merely an example of a multimodal construction site restricted area fence alarm device based on genetic algorithms, and does not constitute a limitation of the multimodal construction site restricted area fence alarm device based on genetic algorithms. It may include more or fewer components than shown in the figure, or combine certain components, or different components. For example, the multimodal construction site restricted area fence alarm device based on genetic algorithms may also include input / output devices, network access devices, buses, etc.

[0077] The processor 81 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0078] The memory 82 can be an internal storage unit of the multimodal construction site restricted area fence alarm device based on genetic algorithms, such as the hard disk or memory of the device. The memory 82 can also be an external storage device, such as a plug-in hard disk, SmartMedia Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the device. Furthermore, the memory 82 can include both internal storage units and external storage devices. The memory 82 is used to store the computer program and other programs and data required by the multimodal construction site restricted area fence alarm device. The memory 82 can also be used to temporarily store data that has been output or will be output.

[0079] Example 3:

[0080] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described in Embodiment 1.

[0081] The computer-readable medium shown can be any means that can contain, store, communicate, propagate, or transmit a program for use in or in conjunction with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, for example, by optically scanning the paper or other medium, then editing, interpreting, or otherwise processing it as necessary to obtain the program electronically, and then storing it in computer memory.

[0082] The above embodiments are merely illustrative of the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made based on the essence of the content of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A genetic algorithm-based multi-modal construction site no-entry zone fence warning method, characterized by, The method comprises: inputting a plan of a construction site, marking construction site information; dividing the plan into a plurality of grid cells, each grid cell being set as a fence or blank; identifying targets in the fence area, using a genetic algorithm to extract target features to identify whether a target has entered the fence area, and issuing an alarm prompt if a target has entered; the use of a genetic algorithm to extract target features includes: initializing a population, each individual representing a possible target feature; defining a fitness function for evaluating the genetic information characteristics of each individual; calculating the fitness value of each individual according to the fitness function, and sorting the individuals from high to low according to the fitness value; according to the preset crossover probability and mutation probability, the individuals in the population are subjected to crossover and mutation operation to generate new individuals; select two individuals as parents according to the probability of the fitness value, then perform single-point crossover at a random position to generate two offspring; each bit of each offspring is subjected to mutation operation; add the newly generated N offspring to the population, and eliminate the N individuals with the lowest fitness value to keep the population size unchanged; N is a positive integer; determine whether the preset termination condition is reached, and output the individual with the highest fitness value in the current population as the optimal feature if the termination condition is reached; the fitness function is: f(x) = R_1 * I_1(x) + R_2 * I_2(x) + R_3 * I_3(x) - C_1 * L(x) - C_2 * S(x) where x represents a fence scheme; R_1, R_2, R_3 represent the reward value when the constraint condition is met; C_1, C_2 represent the coefficients of the length and shape of the fence; I_1(x), I_2(x), I_3(x) represent whether the constraint condition is met, and if it is met, it is 1.

2. The genetic algorithm-based multi-modal construction site no-entry zone fence warning method of claim 1, wherein, The construction site information includes dangerous areas, entrances and exits, and pedestrian density.

3. The genetic algorithm based multi-modal construction site no-go zone fencing alerting method of claim 1, wherein, The target recognition in the fence area includes face recognition, pedestrian recognition, and vehicle recognition.

4. The genetic algorithm-based multi-modal construction site no-entry zone fence warning method of claim 3, wherein, The face recognition includes: using a genetic algorithm to extract local features of a face, using eyes and forehead to verify, and then using other local features to assist in face verification.

5. The genetic algorithm based multi-modal construction site no-entry zone fencing alerting method of claim 3, wherein, The pedestrian recognition includes: extracting pedestrian features using a genetic algorithm to extract robust pedestrian features for recognition.

6. The genetic algorithm based multi-modal construction site no-entry zone fencing alerting method of claim 3, wherein, The vehicle recognition includes: extracting vehicle features in the construction site using a genetic algorithm to extract robust vehicle features for recognition. 7.A genetic algorithm based multi-modal construction site no-entry zone fence warning device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the steps of the method according to any one of claims 1 to 6.

8. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 7. The computer program is executed by the processor to realize the steps of the method according to any one of claims 1 to 6.

Images