CN121259672B - Dangerous rock rapid screening and priority review indicating system based on unmanned aerial vehicle image - Google Patents
Dangerous rock rapid screening and priority review indicating system based on unmanned aerial vehicle imageInfo
- Publication number
- CN121259672B CN121259672B CN202511825051.2A CN202511825051A CN121259672B CN 121259672 B CN121259672 B CN 121259672B CN 202511825051 A CN202511825051 A CN 202511825051A CN 121259672 B CN121259672 B CN 121259672B
- Authority
- CN
- China
- Prior art keywords
- risk
- indicates
- cloud
- rock
- review
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the technical field of geological disaster monitoring, and particularly relates to a dangerous rock rapid screening and priority review indicating system based on unmanned aerial vehicle images; the cloud analysis module is a ground server or a remote cloud server, the cloud analysis module and the unmanned aerial vehicle platform transmit data and instructions through a wireless communication link, the cloud analysis module is used for comprehensively analyzing the rock wall screening result and historical data, generating a review plan and instructions and transmitting the review plan and instructions back to the unmanned aerial vehicle platform or a ground station for execution and scheduling, and the cloud analysis module is used for obviously improving identification timeliness and scheduling scientificity through edge cloud cooperation and review closed loops.
Description
Technical Field
The invention belongs to the technical field of geological disaster monitoring, and particularly relates to a dangerous rock rapid screening and priority review indicating system based on unmanned aerial vehicle images.
Background
In the field of geological disaster prevention and control of mountain slopes and the like, dangerous rock (potentially unstable rock mass with collapse conditions and precursor appearance) forms serious threat to personnel and facilities below. Traditionally, dangerous rock identification and stability assessment mainly rely on manual field investigation, namely engineering geologists need to deeply survey and draw information such as the position, scale, fracture occurrence and the like of the dangerous rock on site, and the stability of the rock mass is judged by combining experience. However, for high-level steep slopes, manual climbing measurement is very difficult, and key information of a dangerous rock structure surface cannot be comprehensively obtained. This results in a large error in the evaluation of the stability of the dangerous rock, which in turn affects the pertinence and timeliness of the abatement measures. In addition, manual investigation itself presents a safety hazard. Therefore, under complex terrains such as high and steep slopes, the unmanned aerial vehicle is introduced to carry out non-contact dangerous rock inspection, and the unmanned aerial vehicle is provided with sensors such as a high-definition camera and the like, so that images and three-dimensional information of dangerous rock can be acquired at a safe distance, the limit that personnel cannot reach is overcome, and a new means is provided for dangerous rock monitoring.
Currently, unmanned aerial vehicles have been applied to dangerous rock inspection and data acquisition. However, in the prior art, unmanned aerial vehicle dangerous rock patrol has the following disadvantages:
The manual screening efficiency is low, a large number of images acquired in the inspection process are usually required to be manually checked and screened frame by frame after flying, and hidden dangers are difficult to discover from massive images in time. Hundreds of slope photos can be acquired in one unmanned aerial vehicle flight, manual analysis is time-consuming, and is easily influenced by subjective factors, so that important signs can be omitted.
High risk area image data accumulation, namely in the face of massive data accumulated by frequent inspection, the existing means lack effective on-site real-time processing, and the data are often detained in an unmanned plane or a storage device to wait for offline analysis. The potential dangerous rock hidden danger cannot be positioned rapidly due to the hysteresis, and the timeliness of early warning is reduced.
And the existing inspection mode usually judges and arranges subsequent inspection according to a fixed period or experience, and lacks scientific quantitative indexes based on risk degree and development trend. This may lead to situations where truly high risk rock mass fails to review in time, missing the best opportunity for intervention, frequently inspecting areas where changes are insignificant, and causing waste of manpower and material resources.
In summary, a new method is urgently needed in the current unmanned dangerous rock inspection process, the on-site data screening efficiency can be improved, high-risk rock mass can be timely identified, objective basis is provided for subsequent re-inspection, and therefore the overall effect of dangerous rock monitoring and early warning is improved.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a dangerous rock rapid screening and priority review indicating system based on an unmanned aerial vehicle image, which comprises an unmanned aerial vehicle platform, an end side computing module and a cloud analyzing module;
the unmanned aerial vehicle platform is used for flying according to a preset routing inspection route, continuously collecting rock wall images along the flying path and transmitting the rock wall images to the end side computing module;
The end side computing module is arranged on the unmanned aerial vehicle platform and is used for processing the rock wall image while the unmanned aerial vehicle platform is in flight inspection to obtain a rock wall screening result;
The cloud analysis module is a ground server or a remote cloud server, the cloud analysis module and the unmanned aerial vehicle end transmit data and instructions through a wireless communication link, the cloud analysis module comprehensively analyzes the screening result and the historical data, generates a review plan and instructions, and transmits the review plan and instructions back to the unmanned aerial vehicle platform or the ground station for execution and scheduling, and closed-loop operation is performed.
The unmanned aerial vehicle platform comprises an aircraft body, a flight controller, a positioning module, a data communication unit and a camera module, wherein the unmanned aerial vehicle platform is used for continuously collecting rock wall images along the flight path, the camera module is used for capturing the rock wall images according to the fixed aerial belt overlapping rate, each frame of image is provided with a time stamp and geographic coordinate and gesture information acquired from the positioning module, the camera module is used for transmitting the rock wall images to an end side computing module, and the data communication unit is used for transmitting the processing results of the end side computing module to a cloud analyzing module.
Preferably, the process of processing the rock wall image by the end side computing module comprises the following steps:
the rock wall image is preprocessed, including distortion correction, illumination balance and noise filtration, ROI screening is adopted, and the rock wall image is focused on the exposed part of the rock in the image to obtain the preprocessed rock wall image;
processing the preprocessed rock wall image by adopting an embedded deep learning model to obtain dangerous rock symptom characteristics;
Calculating a risk score according to the preset feature weight and the dangerous rock symptom feature;
and taking the rock wall image and the corresponding dangerous rock symptom characteristics, risk scores and risk grades as risk rock wall screening results.
Further, the formula for calculating the risk score is expressed as:
wherein, Representing a risk score that is indicative of the risk,Indicating the density of the cracks and,The degree of penetration is indicated by the degree of penetration,Indicating the degree of disadvantage of the camber,Indicating that the connectivity is to be made to the air,Represents the parallel length ratio of the downhill slope,Representing the false detection degree of vegetation;、、、、 And Respectively, the weights of the corresponding features.
The cloud analysis module comprises a data receiving and storing unit, a risk summarizing and evaluating unit, a history change analysis unit, a review priority decision unit and a task scheduling and issuing unit;
The data receiving and storing unit is used for receiving the rock wall screening result and writing the rock wall screening result into the cloud database, and establishing record entries for each unmanned aerial vehicle task;
The history change analysis unit is used for calling history item data and judging the risk development trend of the risk area according to the history item data;
the risk summarizing and evaluating unit is used for updating the cloud database according to the risk development trend of the risk area;
the review priority decision unit generates a priority list according to the latest cloud database information;
the task scheduling and issuing unit generates a review plan according to the priority list, and displays the review plan or directly sends a review instruction to the unmanned plane platform.
Further, the process of updating the cloud database includes:
Comparing the current risk area with the risk area in the original cloud database, judging whether the current risk area is the new risk area, if so, recording dangerous rock symptom characteristics of the corresponding risk area in the cloud database, otherwise, updating the dangerous rock symptom characteristics of the corresponding risk area in the cloud database;
and acquiring a corrected risk parameter, and correcting the risk score and the risk grade of the risk area according to the dangerous rock symptom characteristics of the risk area and the corrected risk parameter.
Further, the formula for correcting the risk score of the risk area is expressed as:
wherein, Representing the risk score after the correction,The intermediate parameter is represented by a value representing,The trend score is indicated as such,A three-dimensional consistency score is represented,The scene criticality is represented by the term,The quality consistency is indicated by the fact that,Representing stable evidence;、、、 And Respectively the weights of the corresponding features are represented,Representing the original risk score of the subject,Representation ofA function.
Preferably, the formula for generating the priority list by the review priority decision unit is expressed as:
)
wherein, The priority value is indicated as such,The latest risk score is indicated,Indicating that the facility is in proximity to the weight,Represents a risk development trend factor and is used for the development of the risk,The time interval factor is represented as such,Indicating the time to date of the last review,Representing the normalized dimension of the scale of the sample,Indicating the set upper limit of the patrol interval,Representation ofThe function of the function is that,Representing a minimum value;、 And Risk weight, facility weight, and time weight, respectively.
The system uses a high risk triggering and energy self-adaptive strategy, and comprises the steps of controlling an unmanned aerial vehicle platform to hover and take multiple images when a risk score calculated by an end side calculation module or a priority value calculated by a cloud analysis module is larger than a preset threshold, only reporting a structuring result and a thumbnail when a link is limited, returning a large image after the unmanned aerial vehicle platform returns, buffering records which are reported by failure, carrying out batch replenishment after returning, entering an accurate photographing mode when the electric quantity of the unmanned aerial vehicle platform is lower than a first electric quantity threshold, and returning the unmanned aerial vehicle platform when the electric quantity is lower than a second electric quantity threshold.
The beneficial effects of the invention are as follows:
The method greatly improves the field recognition efficiency, can screen images in real time through an end-side intelligent algorithm in the cruising process, reduces manual intervention of a large amount of risk-free data, achieves early discovery and early marking of hidden danger, optimizes data transmission and processing, filters and compresses information quantity at the source by edge calculation, only reports important risk data, reduces wireless transmission pressure and cloud computing burden, realizes high-efficiency data utilization, realizes risk grading management, innovatively introduces dangerous rock risk scoring and grading mechanisms, enables quantitative evaluation of hidden danger points at each place, and is convenient for management according to the severity of dangerous situations;
The cloud provides objective review sequence suggestions according to risk classification and evolution trend, overcomes blindness of past experience decision, and ensures timely review of high-risk points without wasting energy on low-risk points;
the system forms a complete closed loop of 'inspection screening-risk assessment-review guidance', can remarkably enhance the initiative and timeliness of geological disaster monitoring, and plays a key role in early warning and prevention of dangerous rock collapse.
Drawings
FIG. 1 is a block diagram of a rapid screening and first review indicator system for dangerous rock in the present invention;
FIG. 2 is a data flow chart of a quick screening and priority review indication system for dangerous rock in the invention;
fig. 3 is a flow chart of a task performed by the unmanned aerial vehicle in the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The invention provides a dangerous rock rapid screening and priority review indicating system based on an unmanned aerial vehicle image, which is shown in fig. 1 and comprises an unmanned aerial vehicle platform, an end side computing module and a cloud analyzing module, wherein the end side computing module is arranged on the unmanned aerial vehicle platform and is connected with a camera shooting/positioning/data communication unit, a processed result is sent to the cloud analyzing module through a wireless link by the data communication unit, and the cloud transmits a command and a plan back to an unmanned aerial vehicle/ground station to form a closed loop with edge-cloud cooperation.
The unmanned aerial vehicle platform is used for flying according to a preset inspection route, continuously collecting rock wall images along the flying path and transmitting the rock wall images to the end side computing module, acquiring rock wall screening results processed by the end side computing module and transmitting the rock wall screening results to the cloud analysis module, wherein the unmanned aerial vehicle platform comprises the following specific components:
The unmanned plane platform consists of an aircraft body, a flight controller (collectively called 'unmanned plane'), a positioning module (an inertial navigation unit (IMU/INS) +a satellite navigation positioning module (GNSS, such as GPS/Beidou)), a data communication unit (a data transmission/image transmission/control link) and a camera module (at least one high-resolution camera) through fusion and calculation of the output position and the gesture.
The unmanned aerial vehicle preferably adopts a multi-rotor unmanned aerial vehicle and has hovering and low-speed cruising capabilities so as to observe the slope of the mountain in a short distance. The unmanned plane platform continuously collects rock wall images along the flight path, specifically comprises the steps that an imaging module carries out aerial photography according to a fixed aerial zone overlapping rate, rock wall images are captured, each frame of image is bound with geographic coordinates and attitude information (longitude and latitude, elevation and attitude angle (roll/pitch/course)) output by a positioning module, and camera sampling moments are aligned with IMU/INS+GNSS resolving results according to a unified time standard (PPS/timestamp), and metadata are written for subsequent space positioning and span period comparison. The method and the device ensure that the position of each suspected dangerous rock is positioned later, spatial references are provided for comparing changes of different times, images acquired by the camera module are transmitted into the end side computing module through the built-in bus to be processed in real time, and the processed key results and necessary original pictures are reported to the cloud analyzing module through the data communication unit. The unmanned plane platform is designed to pay attention to stability and safety, wherein an aircraft is provided with an obstacle avoidance sensor and a redundant control algorithm is executed by a flight control, and when the unmanned plane platform flies in a steep area, the obstacle avoidance sensing and the flight control limiting strategy are combined to avoid terrain collision, and the stability of images is ensured to meet the algorithm requirement.
The end side computing module is arranged on an unmanned aerial vehicle platform, such as an embedded AI computing board card (with a GPU (graphic processing unit) accelerating function) or a high-performance flight control integrated machine, deploys a dangerous rock identification algorithm specially optimized for edge equipment, and comprises a lightweight deep learning model and a rule-based evaluation program, and is used for processing a rock wall image while the unmanned aerial vehicle platform is in flight inspection to obtain a rock wall screening result. Specific:
And each time a new image is acquired, the terminal side computing module immediately analyzes the new image in real time, and the process is as follows:
And preprocessing the rock wall image, including distortion correction, illumination equalization and noise filtration, so as to improve the accuracy of key feature extraction. Preferably, ROI (region of interest) screening is then used to automatically focus on the exposed rock portions in the frame, thereby reducing interference in unrelated areas such as vegetation, sky, etc.
Processing the preprocessed rock wall image by adopting an embedded deep learning model to obtain dangerous rock symptom characteristics, and specifically:
the embedded deep learning model can automatically detect information such as cracks, collapse precursor signs and overhanging degree of a rock face through a large number of sample training aiming at dangerous rock characteristics. Specifically, the algorithm firstly extracts a fracture network on a rock wall image, and calculates six normalized features, namely dangerous rock sign features, in each candidate risk area, namely (1) the fracture density D, namely the total length of a fracture skeleton in the area and the area The ratio of areas (or number of strips per unit area) and is linearly normalized to [0,1]. (2) And the penetration degree C is normalized to [0,1] whether the crack skeleton forms a communication path crossing the upper/lower edge or the left/right edge of the region or not and the crossing proportion (the ratio of the maximum communication component length to the characteristic dimension of the region) of the communication path. (3) Camber disfigurement O, estimating camber/suspension degree (such as front edge bulge, bottom crack and the like) of rock mass by texture and geometric form, and normalizing the camber angle or suspension index to [0,1] according to a threshold value. (4) The void connectivity F is normalized to [0,1] by the contact or neighbor ratio of the crack end point/connected component to the void boundary (cliff edge, free face). (5) Parallel length of downhill is equal to the ratio S: major direction and slope direction of cracks the included angle is smaller than the set threshold valueThe ratio of the skeleton length of the crack to the total length is normalized to [0,1]. (6) The false detection degree V of vegetation is the proportion of pixels which are shielded by vegetation or are misjudged to be cracks and are obtained by vegetation discrimination (such as color/texture/NIR indexes), and the proportion is normalized to [0,1].
And then calculating a risk score according to the preset feature weight and the dangerous rock symptom feature, wherein a formula for calculating the risk score is expressed as follows:
wherein, Representing a risk score that is indicative of the risk,Indicating the density of the cracks and,The degree of penetration is indicated by the degree of penetration,Indicating the degree of disadvantage of the camber,Indicating that the connectivity is to be made to the air,Represents the parallel length ratio of the downhill slope,Representing the false detection degree of vegetation;、、、、 And Respectively, the weights of the corresponding features.
The risk level of the rock wall image is classified according to the risk score, preferably, the rock wall image is classified into a high level, a middle level and a low level according to a preset threshold, wherein the score is higher than the threshold L1 and is judged to be high risk (class I), important attention is required, the rock wall image is classified into a medium risk (class II) between the threshold L1 and the threshold L2, and the rock wall image is classified into a low risk (class III) below the threshold L2.
The weight and the threshold can be automatically calibrated according to the scene, but the weight E [0.05,0.40] and the threshold E [0.30,0.80] are required to be within the range.
For images with very low risk and no obvious cracks, the modules can be marked directly as risk-free. The terminal side screening process is fast and efficient, and unmanned aerial vehicle flight inspection and data analysis can be synchronously carried out. For example, dangerous rock identification and scoring of the image is completed within a few seconds of the unmanned hovering to take a picture of a rock wall. The first-round risk rock wall screening result is generated by the end side computing module when the unmanned aerial vehicle does not leave the scene, wherein the risk rock wall screening result comprises a rock wall image, corresponding dangerous rock symptom characteristics, risk scores and risk grade information.
In some preferred embodiments of the invention, a rule-based expert system may also be used to rank risk, for example, if multiple vertical through cracks are found and the rock mass is tilted forward, a high risk is determined.
Through the design of the end side module, the invention realizes the preliminary screening and quantitative evaluation of dangerous rock hidden danger on the end side (on-board of the unmanned aerial vehicle). The on-site instant screening obviously shortens the time from data acquisition to risk discovery, avoids the backlog of inspection data, and wins precious time for subsequent links. In addition, because a large number of safety area images are screened, only key risk information is reserved for uploading, and communication and cloud processing pressure is greatly reduced.
The cloud analysis module is a ground server or a remote cloud server, and is used for transmitting data and instructions with the unmanned aerial vehicle platform through a wireless communication link, comprehensively analyzing the rock wall screening result and the historical data, generating a review plan and instructions and transmitting the review plan and instructions back to the unmanned aerial vehicle platform or the ground station for execution and scheduling, and specifically:
And after receiving the data from the unmanned aerial vehicle inspection, the cloud analysis module executes deeper analysis and decision generation, and the workflow of the cloud analysis module is shown in figure 2. The cloud analysis module comprises a data receiving and storing unit, a risk summarizing and evaluating unit, a history change analysis unit, a review priority decision unit and a task scheduling and issuing unit.
The data receiving and storing unit is used for receiving the rock wall screening result and writing the rock wall screening result into the cloud database, and establishing record entries for each unmanned aerial vehicle task, wherein the entry data preferably comprises task time, risk areas and suspected dangerous rock lists identified by the end sides.
The risk area refers to a high risk space unit which is aggregated by an algorithm in space and is stored in a polygon (WGS-84 coordinate) or grid index, and the geometric information (polygon, circumscribed rectangle and area) of the high risk space unit, the average S ̄/maximum risk score S_max in the area, an image slice/tile list and basic topography attributes (average gradient/slope direction) are recorded.
And the suspected dangerous rock list identified by the end side refers to a candidate body set (rock bodies with high and medium risk grades) output by the end side in real time in the task process. Each candidate body comprises a unique ID, spatial positioning (longitude and latitude/elevation/corresponding image fragment index), model confidence, risk score S, risk parameter vector (dangerous rock symptom feature), geometric statistics (total length of crack, maximum width, number of connected components, main direction), three-dimensional terrain attribute (local gradient, slope direction, shortest distance from the free surface) and the like.
The cloud database organizes data by geographic space index, so that observations of different times at the same place can be stored in an associated mode, and history comparison is facilitated.
The history change analysis unit is used for calling history item data and judging the risk development trend of the risk area according to the history item data:
For each identified dangerous rock in the cloud database, the system invokes its characterization data (record over time of fracture density, length, risk score, etc.) for a historical survey. The latest data are compared with the data before the last time or times through an algorithm, and the difference of the previous results, such as how much the crack length is increased, how many cracks are newly appeared, how much the risk score is increased/decreased, and the like, is calculated.
The combination of the multi-period data can also judge the change trend, namely, if the score of a certain rock body continuously rises twice, the trend is worsened, and if the score is kept stable for a plurality of times, the trend is stable. The cloud generates a trend evaluation result (trend label) for each region (e.g., a "continuous rise in risk", "substantially stable" or "slow down", etc.). The evolution analysis provides scientific basis for review decision-compared with single result, the evolution analysis can reflect the development dynamics of hidden danger.
The risk summarizing and evaluating unit is used for updating the cloud database according to the risk development trend of the risk area. The process of updating the cloud database comprises the following steps:
Comparing the current risk area with the risk area in the original cloud database, judging whether the current risk area is the new risk area, if so, recording dangerous rock sign features of the corresponding risk area in the cloud database, and otherwise, updating the dangerous rock sign features of the corresponding risk area in the cloud database.
Next, the cloud performs fine-grained evaluation for each risk area, and more complex algorithms and models can be invoked to verify and supplement end-side results using the stronger computing power of the cloud. Specific:
And acquiring a risk correction parameter, and correcting a risk score and a risk grade of the risk area according to the dangerous rock symptom characteristics and the risk correction parameter of the risk area. Specific:
For high risk points, the cloud can extract multi-angle images or three-dimensional point cloud data (even laser point cloud if the unmanned aerial vehicle is provided with multi-view images), and the specific instability mode and the risk degree of the rock mass block are estimated through structural plane analysis, three-dimensional reconstruction or finite element stability simulation and other means. These advanced analyses help to reduce false positives and improve accuracy. For example, the end side may judge high risk because of crack shadow, but the cloud may adjust the risk rating down after confirming the crack depth through multiple views, and vice versa, the cloud may analyze the hidden danger details not detected by the end side finely, so as to adjust the risk rating up.
For specific quantitative adjustment, a new primary risk score is obtained according to dangerous rock symptom characteristics of a risk areaThe formula is as described above. Then, acquiring risk correction parameters including trend score, three-dimensional consistency score, scene criticality, quality consistency and stable evidence according to the risk correction parameters and original risk scoreAnd correcting to obtain a new risk score. The risk scoring formula for correcting the risk area designed by the invention is as follows:
wherein, Representing the risk score after the correction,Representing intermediate parameters, T representing trend score, score slope from last multiple observationsNormalization results in: the value range is [ -1,1]. For normalizing the dimensions, preference is given to=0.06/Week (refer to every past 1 week, rising by 0.06). The larger the slope (the larger the absolute value), the larger the amplitude of the trend component.Represents three-dimensional consistency scores (multi-angle images or point clouds confirm consistency of camber and connectivity),Representing scene criticality (distance weight from targets such as roads/pipelines/residences),Quality consistency (multi-model consistency and manual review consistency plus score) is represented,Indicating evidence of stability (no change over time or signs of reverse recovery withhold);、、、 And Respectively the weights of the corresponding features are represented,Representing the original risk score of the subject,Representation ofA function.
M can calculate consistency scores by comparing geometric features of cracks extracted from a multi-view image (or three-dimensional point cloud) of the same candidate, K can be mapped into weight values (the weight is larger when the distance is closer) in a [0,1] interval according to actual distances between a rock body and a key target (such as a road, a pipeline, a house and the like) by a linear or piecewise function, Q can be given by counting consistency rates of a plurality of detection models and manual labeling results (for example, if both models and manual judgment are in a crack, Q can take a high value), and R can give positive/negative values according to a trend of historical risk score changes (if scores are not obviously increased, stable evidence is judged to be buckled).
The function is expressed as:
and (5) defining the risk grade again according to the updated risk score.
And the review priority decision unit generates a priority list according to the latest cloud database information. Specific:
The cloud gives a review priority sequence to dangerous rock points to be observed according to a certain strategy, wherein the strategy comprehensively considers factors such as risk level (I level high risk is higher than II level medium risk, II level is higher than III level low risk), recent change amplitude (recent worsening is higher than unchanged), last inspection time (the longer the interval is, the faster the review should be), and the field environment (for example, hidden danger adjacent to important facilities should be higher priority) needs to be considered if necessary. Through multi-factor decision making, the system generates a weighted priority list. For example, if a hidden danger is a high risk before and the score continues to rise, the priority is highest, otherwise, if a medium risk point which is always present is not changed obviously for a long time, the recheck urgency can be reduced properly, and in order to quantify the priority specifically, the invention designs a formula for generating a priority list:
wherein, The priority value is indicated as such,The latest risk score is indicated,Indicating that the facility is in proximity to the weight,Represents a risk development trend factor and is used for the development of the risk,The time interval factor is represented as such,Indicating the time to date of the last review,Representing the normalized dimension of the scale of the sample,Indicating the set upper limit of the patrol interval,Representation ofThe function of the function is that,Representing the minimum.、AndRisk weight, facility weight and time weight, respectively, preferably,=0.5, B=0.3, c=0.2. S is in the range of [0,1], H is in the range of [0,1],Is in the range of [ -1,1],The range of (2) is [0,1], and therefore P takes on the value [0, 2]. Preferably, the facility proximity weights (between 0 and 1) can be valued by key facilities (school/hospital/rail/high speed/dense resident) 30m 1.0, 30-60 m 0.8, 60-100 m:0.5;100 m:0
Important facilities (province/county roads/pipelines/scenic areas roads/base stations) are not more than 20 m:0.8, 20-50 m:0.6, 50-80 m:0.3 and 80 m:0
Common facilities (rural roads/power transmission towers/scattered households/farmland) are less than or equal to 20m to 0.5, 20-40 m to 0.3, and 40 m to 0 is added with 0.1 along the slope, so that effective protection is reduced by 0.2.
In particular, the method for value-taking in the case of simultaneous proximity to a plurality of facilities, i.e. the presence of an overlap, is as follows:
The greatest weight in the corresponding values of all categories that overlap (i.e., the type of facility deemed most sensitive to the proximity of the hazard) is employed to ensure conservation of risk assessment. For example, if a hazard point is located in both the critical facility-high speed (30 m) vicinity (weight 1.0) and the critical facility-pipeline (20-50 m) vicinity (weight 0.6), then weight 1.0 may be taken as the integrated facility proximity weight.
The segmentation is represented as follows:
specific priority levels can be generated according to the priority values and preset thresholds, for example, P is more than or equal to 1.20, immediate review is performed, P is more than or equal to 0.80 and less than or equal to P is less than 1.20, next voyage priority is performed, and P is less than or equal to 0.80, and routine inspection is performed.
The task scheduling and issuing unit generates a review plan according to the priority list, and displays the review plan or directly sends a review instruction to the unmanned plane platform. Specific:
the cloud converts the prioritized review suggestions into specific inspection tasks. For example, a high priority list of points may form the key coordinates for the next sub-drone route plan, with the system automatically generating the review plan. These plans are issued to the unmanned aerial vehicle ground station or control center via the network, prompting the operator or autonomous unmanned aerial vehicle to perform targeted review cruising. For significant potential hazards requiring immediate review, the system may send an early warning notice and suggest immediate dispatch of the drone or manual forward verification and handling. Accordingly, the low priority areas are arranged for attention in the routine inspection. The task issuing can be fully automatic (in the unmanned aerial vehicle cluster system, the cloud directly sends a review instruction to the unmanned aerial vehicle platform to instruct the idle unmanned aerial vehicle to go to a high-priority point for review), and a review plan can be displayed to a manager through a platform interface to prompt, so that decision-making is adopted and executed. Regardless of the way, a set of review management mechanism driven by risk ordering is finally established, so that limited patrol resources are optimally distributed, namely, the safety of high-risk points is guaranteed, and then general patrol is considered.
The dangerous rock rapid screening and priority review indicating system designed by the invention uses a high risk triggering and energy self-adaptive strategy, and comprises the steps of controlling an unmanned aerial vehicle platform to hover and supplement a plurality of images when a risk score calculated by an end side calculating module or a priority value calculated by a cloud analyzing module is larger than a preset threshold, only reporting a structuring result and a thumbnail when a link is limited, returning a large image after the unmanned aerial vehicle platform returns to the home position, caching records reported by failure of the end side calculating module, supplementing in batches after returning to the home position, entering a fine photographing mode when the electric quantity of the unmanned aerial vehicle platform is lower than a first electric quantity threshold, and returning the unmanned aerial vehicle platform when the electric quantity is lower than a second electric quantity threshold.
In conclusion, through the cooperation of the units, the invention realizes the secondary screening and overall planning of the screening result of the opposite end side, truly falls the thought of screening first and then rechecking to the reality, and the system designed by the invention obviously improves the identification timeliness and the scheduling scientificity through the side cloud cooperation and rechecking closed loop, and has good application prospect.
While the foregoing is directed to embodiments, aspects and advantages of the present invention, other and further details of the invention may be had by the foregoing description, it will be understood that the foregoing embodiments are merely exemplary of the invention, and that any changes, substitutions, alterations, etc. which may be made herein without departing from the spirit and principles of the invention.
Claims (5)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202511825051.2A CN121259672B (en) | 2025-12-05 | 2025-12-05 | Dangerous rock rapid screening and priority review indicating system based on unmanned aerial vehicle image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202511825051.2A CN121259672B (en) | 2025-12-05 | 2025-12-05 | Dangerous rock rapid screening and priority review indicating system based on unmanned aerial vehicle image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN121259672A CN121259672A (en) | 2026-01-02 |
| CN121259672B true CN121259672B (en) | 2026-04-10 |
Family
ID=98195926
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202511825051.2A Active CN121259672B (en) | 2025-12-05 | 2025-12-05 | Dangerous rock rapid screening and priority review indicating system based on unmanned aerial vehicle image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN121259672B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116661489A (en) * | 2023-05-19 | 2023-08-29 | 浙江华络通信设备有限公司 | Mine inspection drone and its control method based on 5G independent networking |
| CN118762479A (en) * | 2024-09-06 | 2024-10-11 | 高精特(成都)大数据科技有限公司 | A rock and soil slope geological disaster early warning system and method |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119251714A (en) * | 2024-09-29 | 2025-01-03 | 西南交通大学 | Risk assessment method for dangerous rock mass based on crack density analysis |
| CN120375040B (en) * | 2025-03-21 | 2025-10-28 | 广西壮族自治区地质环境监测站 | Dangerous rock early warning method, system, equipment and medium |
-
2025
- 2025-12-05 CN CN202511825051.2A patent/CN121259672B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116661489A (en) * | 2023-05-19 | 2023-08-29 | 浙江华络通信设备有限公司 | Mine inspection drone and its control method based on 5G independent networking |
| CN118762479A (en) * | 2024-09-06 | 2024-10-11 | 高精特(成都)大数据科技有限公司 | A rock and soil slope geological disaster early warning system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN121259672A (en) | 2026-01-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114373138B (en) | A high-speed railway fully automatic drone inspection method and system | |
| CN118608709A (en) | Real-time 3D modeling of earthwork engineering, calculation of earthwork quantity and construction progress monitoring method based on drone measurement | |
| CN120431683B (en) | River Basin Disaster Monitoring and Warning Method and System Based on Multi-source Monitoring Data Analysis | |
| CN103984355A (en) | Routing inspection flying robot and overhead power line distance prediction and maintaining method | |
| CN111339826B (en) | A linear sensor network framework detection system for landslide UAV | |
| CN119440085B (en) | A highway bridge slope maintenance drone inspection method and system | |
| CN118170164A (en) | A photovoltaic power station construction inspection method | |
| CN119672566A (en) | A highway waterlogging detection and early warning method based on drones | |
| CN119245623A (en) | Method, device and system for updating terrain data | |
| CN116363530B (en) | Method and device for positioning expressway pavement diseases | |
| CN120635756B (en) | Intelligent identification method, system and device for abnormal state of transformer | |
| CN120932100B (en) | A System and Method for Identifying Landslide Hazardous Areas of Embankments Based on UAV and Remote Sensing Imagery | |
| CN121259672B (en) | Dangerous rock rapid screening and priority review indicating system based on unmanned aerial vehicle image | |
| CN120411223A (en) | An efficient method for positioning photovoltaic inspection results using drones | |
| CN118334543B (en) | Unmanned aerial vehicle-based well lid identification method and device, computer equipment and medium | |
| CN120213004A (en) | A method for generating railway electronic map | |
| CN119814972A (en) | An open-pit mine information monitoring system | |
| CN114659499B (en) | Smart city 3D map model photography establishment method based on unmanned aerial vehicle technology | |
| CN118758268A (en) | Railway status diagnosis method based on 2D/3D remote sensing of UAV at low altitude | |
| CN118038293A (en) | Real-time inspection device, method and system for road surface diseases based on unmanned aerial vehicle | |
| CN121527550B (en) | Contact net foreign matter and steel rail damage detection method based on unmanned aerial vehicle binocular vision | |
| CN119810517B (en) | Wind power plant overhead line defect intelligent identification system based on unmanned aerial vehicle autonomous inspection | |
| CN121390662A (en) | Surface mine space-earth integrated management method and system based on unmanned airport | |
| CN117475337A (en) | A device and method for rapid identification and evaluation of coal gangue piles in Yunnan-Guizhou karst mountainous areas | |
| CN121541548A (en) | Intelligent Inspection Control Method and System Based on Image Processing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |