Intelligent hidden danger checking method and system based on image recognition

Abstract

The invention relates to the technical field of intelligent inspection and image analysis, and discloses an intelligent hidden danger checking method and system based on image recognition. The method comprises the following steps: establishing a hidden danger knowledge network for representing hidden danger types, features and cause association relationships; hidden danger images are collected and input into a network during inspection, hidden danger entities are extracted from the hidden danger images, features of the hidden danger entities are matched with hidden danger features in the network, and associated hidden danger types and causes are obtained through network reasoning. And calculating a hidden danger probability according to the association strength of the causes, and generating and outputting an intelligent troubleshooting task bound with the position according to the comparison of the probability and a threshold value so as to guide targeted troubleshooting. And finally, optimizing the association relationship in the hidden danger knowledge network according to field verification feedback for the task. According to the invention, intelligent recognition and cause reasoning of hidden dangers are realized through the knowledge network, the network is optimized by using a service feedback closed loop, and the intelligent depth of hidden danger investigation and the adaptive ability of the system are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent inspection and image analysis, in particular to an intelligent hidden danger investigation method and system based on image recognition. BACKGROUND

[0002] In the daily inspection of the power, chemical industry, construction and other industries, images are collected by cameras or unmanned aerial vehicles, and hidden danger targets are automatically identified by computer vision models, which has become a common technical means to improve inspection efficiency. This kind of conventional technical scheme usually regards hidden danger investigation as a typical target detection or image classification task, and its technical path is to collect labeled data, train an identification model, and deploy the model for prediction. The result output by the model is usually the specific hidden danger object and its position information existing in the image.

[0003] Such conventional technical schemes have limitations. Its function stops at the level of "discovery" and "what is" of hidden dangers, and can only inform that there is a phenomenon or entity that does not conform to safety standards on site. It lacks the ability to analyze the causes behind hidden dangers and cannot answer questions such as "why does this hidden danger occur" or "what deeper faults, management loopholes or environmental factors may cause this phenomenon". This makes the investigation work still highly dependent on expert experience for secondary analysis, and the depth of intelligence is insufficient.

[0004] The existing scheme is an open-loop system. The optimization iteration of the identification model mainly depends on labeling more new samples to improve its recognition accuracy, but the update of the model itself is disconnected from the real results verified in the business closed loop. The system cannot use the feedback information from the on-site verification of front-line personnel to optimize and correct the "association logic between different hidden danger characteristics and specific causes" of deep knowledge. This leads to the static analysis capability of the system, which cannot learn from actual business feedback, and the accuracy and adaptability of its reasoning cannot be enhanced independently along with the use process. SUMMARY

[0005] The purpose of the present application is to provide an intelligent hidden danger investigation method and system based on image recognition to solve the problems raised in the background art.

[0006] To achieve the above purpose, the present application provides an intelligent hidden danger investigation method based on image recognition, which comprises: establishing a hidden danger knowledge network for representing the association relationship between hidden danger types, hidden danger characteristics and hidden danger causes; collecting hidden danger images of target areas on a set inspection path, and inputting the hidden danger images into the hidden danger knowledge network; extracting hidden danger entities existing in the hidden danger images from the initial analysis results output by the hidden danger knowledge network; matching the entity feature of the hidden danger entity with the hidden danger feature in the hidden danger knowledge network, reasoning through the hidden danger knowledge network to obtain a hidden danger type and a hidden danger cause associated with the hidden danger entity; According to the association strength of the hidden danger cause, the hidden danger probability of the corresponding area of the hidden danger image is calculated. According to the comparison result of the numerical value of the hidden danger probability and the preset threshold, the intelligent investigation task bound to the position of the hidden danger image is generated. Output the intelligent investigation task to guide the targeted hidden danger investigation of the target area. According to the field verification feedback of the intelligent investigation task, the association relationship in the hidden danger knowledge network is optimized.

[0007] Preferably, the hidden danger knowledge network for representing the association relationship between hidden danger type, hidden danger feature and hidden danger cause is established, comprising: Obtain historical hidden danger cases, and parse historical hidden danger types, historical image features and historical causes from the historical hidden danger cases; Define the historical hidden danger type as a node type in the hidden danger knowledge network, abstract the historical image feature as an attribute feature in the hidden danger knowledge network, and establish the historical cause as an associated node in the hidden danger knowledge network. Establish a directed connection between the node type, the attribute feature and the associated node, and initialize the weight of the directed connection based on the co-occurrence frequency of the corresponding item in the historical hidden danger case.

[0008] Preferably, the hidden danger entity existing in the hidden danger image is extracted from the initial analysis result output from the hidden danger knowledge network, comprising: Receive the initial analysis result of the hidden danger image from the hidden danger knowledge network, and the initial analysis result contains multiple image regions in the hidden danger image which have potential matching degree with the attribute feature; Frame the entity boundary of the multiple image regions respectively, and enhance the local feature in the region framed by the entity boundary; Identify whether there is an entity feature matching any of the attribute features in the region after the local feature enhancement, and if the matching is successful, define the region framed by the entity boundary as a hidden danger entity.

[0009] Preferably, the matching of the entity feature of the hidden danger entity with the hidden danger feature in the hidden danger knowledge network, the reasoning through the hidden danger knowledge network to obtain the hidden danger type and the hidden danger cause associated with the hidden danger entity, comprising: The entity feature is calculated for similarity with all attribute features in the hidden danger knowledge network, and the attribute feature with the highest similarity is selected as a matching result; A property feature node corresponding to the matching result is located in the hidden danger knowledge network, and all directed connections starting from the property feature node are activated; Path traversal is performed along the activated directed connections, all reachable node types and associated nodes are converged, the converged node types are taken as hidden danger types associated with the hidden danger entity, and the converged associated nodes are taken as hidden danger causes associated with the hidden danger entity.

[0010] Preferably, the hidden danger probability of the hidden danger image corresponding area is calculated according to the association strength of the hidden danger cause, comprising: For each hidden danger cause associated with the hidden danger entity, a path weight from the property feature node corresponding to the matching result to the associated node corresponding to the hidden danger cause is obtained in the hidden danger knowledge network, and the path weight is defined as the association strength of the hidden danger cause; The association strengths of all hidden danger causes associated with the hidden danger entity are weighted and fused to obtain an entity risk value of the hidden danger entity; The sum of the entity risk values of all hidden danger entities in the hidden danger image is counted, and the sum is normalized to map a value between zero and one hundred, and the value is taken as the hidden danger probability of the hidden danger image corresponding area.

[0011] Preferably, after collecting the hidden danger image of the target area, it further comprises: Collect non-image dimension information at the same space-time point as the hidden danger image, the non-image dimension information comprising at least one of environmental temperature information, device vibration information, and gas concentration information; In the hidden danger knowledge network, an auxiliary decision node is established for each type of non-image dimension information; When network reasoning is performed, the quantized value of the non-image dimension information is taken as the input of the auxiliary decision node, and the effective weight of the directed connection from the property feature node to the associated node is adjusted.

[0012] Preferably, the intelligent investigation task bound to the hidden danger image position is generated according to the comparison result of the value of the hidden danger probability and a preset threshold, comprising: When the value of the hidden danger probability is greater than the preset threshold, a task generation instruction is triggered; In response to the task generation instruction, the hidden danger types associated with the hidden danger entity, the hidden danger causes, and the non-image dimension information of the hidden danger image corresponding area are integrated to form a structured description; Generate an intelligent investigation task containing target position, target object and key points of attention based on the position information of the hidden danger image and the structured description.

[0013] Preferably, after outputting the intelligent investigation task, the method further comprises: During the field investigation for the intelligent investigation task, record the on-site disposal measures for the hidden danger entity and the post-disposal state; Compare the on-site disposal measures with the hidden danger causes associated in the hidden danger knowledge network, and mark the root cause node targeted by the on-site disposal measures; Associate the root cause node with the corresponding hidden danger entity to form a verified knowledge association pair.

[0014] Preferably, the method further comprises: Receive the verified knowledge association pair and identify the original directed connection between the root cause node and the attribute feature node; According to the effectiveness verification result of the on-site disposal measures, enhance or weaken the weight of the original directed connection; For redundant connections that exist in the hidden danger knowledge network but have never been associated by the verified knowledge association pair, gradually reduce their weights until they are removed.

[0015] Preferably, the present application further comprises an intelligent hidden danger investigation system based on image recognition, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the intelligent hidden danger investigation method based on image recognition as described above when executing the computer program.

[0016] Compared with the prior art, the present application has the following advantages: By constructing a deep reasoning knowledge network deeply associated with hidden danger types, features and causes, the entity features recognized by image recognition are matched with network nodes, and logical reasoning is performed along the preset association path, realizing the leap from surface phenomenon recognition to deep cause inference. The output of this method is no longer an isolated hidden danger label, but an analysis result containing potential causal chains, so that the hidden danger investigation work can directly focus on the problem root, improving the depth and pertinence of analysis and decision-making.

[0017] By taking the field verification feedback of the intelligent troubleshooting task as the key data, the association and weight between nodes in the knowledge network are continuously optimized, and the system builds a complete closed loop from perception, decision-making to verification and learning. This mechanism makes the reasoning logic of the system no longer static and unchangeable, but can be continuously self-corrected and enhanced according to the feedback of the real scene, so as to gradually improve the accuracy and adaptability of the cause inference in different complex environments.

[0018] The reasoning mechanism of the hidden danger knowledge network converts discrete identification results into explanatory analysis conclusions, providing clear troubleshooting directions for field personnel and reducing their absolute dependence on professional experience. The closed-loop optimization mechanism ensures that the system's analysis ability can evolve continuously with the accumulation of business data, making it a growing intelligent tool that ultimately improves the efficiency and accuracy of hidden danger troubleshooting. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 The working principle diagram of the intelligent hidden danger troubleshooting method based on image recognition described in the present application; Figure 2 The flowchart for establishing the hidden danger knowledge network; Figure 3 The flowchart for matching entity features and performing network reasoning; Figure 4 The comparison chart of effective weights of the knowledge network; Figure 5 The initial knowledge network connection weight distribution chart. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0021] Please refer to Figure 1The application provides an intelligent hidden danger investigation method based on image recognition. The method comprises the following steps: firstly, a hidden danger knowledge network is constructed, which is used to formally represent the internal correlation among hidden danger types, hidden danger characteristics and hidden danger causes. In the application stage, hidden danger images of a target area are collected along a preset inspection path, and the collected hidden danger images are input into the constructed hidden danger knowledge network. The system extracts hidden danger entities actually existing in the images from initial analysis results generated by the hidden danger knowledge network after preliminary analysis of the input images. Then, entity characteristics of the hidden danger entities are matched with standardized hidden danger characteristics stored in the hidden danger knowledge network, and reasoning is performed by using the internal correlation of the knowledge network, so that specific hidden danger types and possible hidden danger causes associated with the hidden danger entities are obtained.

[0022] Based on the correlation strength between each hidden danger cause and the current characteristics inferred, the system calculates a comprehensive hidden danger probability of the area corresponding to the hidden danger image by calculation. The hidden danger probability value is compared with a preset threshold, and an intelligent investigation task bound with geographical location information of the hidden danger image is automatically generated according to the comparison result. The generated intelligent investigation task is output to guide the inspection personnel to perform on-site verification on the target area. Finally, the system iteratively optimizes the internal correlation of the hidden danger knowledge network according to feedback information obtained after on-site verification of the intelligent investigation task, so that the internal correlation is more accurate.

[0023] Embodiment 1: see Figure 2In specific implementations, historical hidden danger cases are acquired, and historical hidden danger types, historical image features, and historical causes are parsed from the historical hidden danger cases. In specific implementations, the historical hidden danger types are defined as node types in the hidden danger knowledge network, the historical image features are abstracted as attribute feature nodes in the hidden danger knowledge network, and the historical causes are established as associated nodes in the hidden danger knowledge network. After parsing the node types, attribute feature nodes, and associated nodes based on the historical hidden danger cases, the abstraction process of the attribute features needs to combine the feature engineering methods in the field of computer vision to perform grayscale, edge detection, and feature point extraction on the historical image features, combine the extracted low-dimensional visual features into a high-dimensional feature vector, and use the high-dimensional feature vector as the core data form of the attribute feature nodes in the hidden danger knowledge network. In addition to being based on the co-occurrence frequency, the initialization and assignment of the directed connection also need to map the weight to the [0, 1] interval through normalization processing to ensure the comparability of the connection weights between different types of nodes. For example, when the historical hidden danger cases have “cable breakage” and “insulation layer cracking image feature” co-occurring 80 times, and the total number of occurrences of “cable breakage” is 100, the initial weight of the directed connection between them is calculated as 0.8, which is in the same quantization dimension as the weight of other associated items. In some embodiments, the directed connections between the node types, attribute feature nodes, and associated nodes are established according to the logical relationships reflected in the historical hidden danger cases, and the initial weight values of the directed connections are initialized and assigned based on the co-occurrence frequencies of the corresponding items in the historical hidden danger cases. It can be understood that the co-occurrence frequency refers to the number of times that the node types, attribute feature nodes, and associated nodes appear simultaneously in the historical hidden danger cases. Optionally, the initial weight of the directed connection is calculated by the formula: ; wherein: represents the initial weight of the directed connection, represents the co-occurrence number of the node types, attribute feature nodes, or associated nodes related to the directed connection in the historical hidden danger cases, represents the total number of occurrences of the starting node of the directed connection in the historical hidden danger cases. In specific implementations, the starting node of the directed connection is a node type or an attribute feature node, and the ending node of the directed connection is an attribute feature node or an associated node. In some embodiments, the parsing process of the historical hidden danger cases includes reading the text records and image data in the case database, and extracting the historical hidden danger types, historical image features, and historical causes. It can be understood that the construction of the hidden danger knowledge network depends on the collection and parsing of the historical hidden danger cases, and the directed connections between the node types, attribute feature nodes, and associated nodes form the associated relationship network of the hidden danger types, hidden danger features, and hidden danger causes.

[0024] Embodiment 2: refer to Figure 3In specific implementations, the initial analysis result generated by the receiving hidden danger knowledge network after processing the input hidden danger image contains multiple image regions in the hidden danger image that have potential matching degrees with the attribute features stored in the hidden danger knowledge network. In specific implementations, the entity bounding box operation is performed on the multiple image regions with potential matching degrees, and image local feature enhancement processing is performed within each region determined by the entity bounding box operation. It can be understood that the local feature enhancement processing aims to highlight the detailed texture and structural information in the image region. When performing the entity bounding box operation on the multiple image regions, the YOLO target detection algorithm is used to generate an initial bounding box, and non-maximum suppression is used to remove redundant boxes with an overlap degree higher than 0.5, thereby ensuring that the bounding box of each potential hidden danger region is unique. The adaptive histogram equalization algorithm is used in the local feature enhancement link to optimize the brightness unevenness and detail blur of the image within the bounding box, thereby improving the matching accuracy of the entity features and the attribute features. The similarity calculation adopts a weighted judgment method combining the cosine similarity and the Euclidean distance, wherein the weight proportion of the cosine similarity is 0.6, and the weight proportion of the Euclidean distance is 0.4. When the comprehensive similarity score is higher than 0.7, it is determined that the matching is successful. For example, the cosine similarity between the “metal corrosion feature vector” of the hidden danger entity and the “rust attribute feature vector” in the knowledge network is 0.75, the Euclidean distance corresponds to a similarity of 0.68, and the comprehensive score is 0.722, which meets the matching threshold requirement. The weight setting is based on the effectiveness statistics of feature matching in historical hidden danger cases. The cosine similarity is more suitable for judging the direction consistency of high-dimensional features, and the Euclidean distance is more suitable for measuring the difference of feature values. The corresponding weight distribution is determined by combining the performance of the two in the past matching tasks. In some embodiments, it is determined whether there is an entity feature that matches any attribute feature in the hidden danger knowledge network in the region after the local feature enhancement processing. If the matching is successful, the region determined by the entity bounding box operation is defined as a hidden danger entity. In specific implementations, the determination basis for the matching success is that the similarity between the entity feature and the attribute feature exceeds the set threshold.

[0025] The implementation process of matching the entity features of the hidden danger entity with the hidden danger features in the hidden danger knowledge network and reasoning through the hidden danger knowledge network. In specific implementations, the extracted entity features of the hidden danger entity are subjected to similarity calculation with all the attribute features stored in the hidden danger knowledge network, and the attribute feature with the highest similarity calculation value is selected as the final matching result. In some embodiments, the similarity calculation can adopt the cosine similarity method, which is expressed by the formula: ; wherein: represents the similarity calculation result, represents the entity feature vector extracted from the hidden danger entity, This represents a feature vector of a certain attribute stored in the hazard knowledge network. The process involves locating the attribute feature node corresponding to the matching result within the hazard knowledge network and activating all directed connections originating from that node. This activation operation implies considering the relationships represented by these directed connections during subsequent reasoning. Optionally, a path traversal is performed along the activated directed connections to aggregate all node types and associated nodes reachable through these connections. The aggregated node types are considered as the hazard types associated with the hazard entity, and the aggregated associated nodes are considered as the hazard causes associated with the hazard entity. In practice, a breadth-first search algorithm is used for node search during path traversal. In practice, the path traversal uses a breadth-first search algorithm to search for nodes. The breadth-first search algorithm starts with the attribute feature node corresponding to the matching result. First, the node is visited and its type is checked. If it is a node type or an associated node, it is added to the aggregation set. Then, all nodes directly connected to the node through directed connections are visited and marked as visited. Then, the next level of connected nodes of these directly connected nodes are visited in the same way to ensure that each node is processed only once, until all nodes reachable through directed connections have been traversed. Finally, the node types accumulated in the aggregation set are used as the hazard types associated with the hazard entity, and the associated nodes are used as the causes of the hazard.

[0026] Example 3: In specific implementation, for each hazard cause associated with a hazard entity, the weight of the path traversed from the attribute feature node corresponding to the matching result to the associated node corresponding to the hazard cause is obtained in the hazard knowledge network. This path weight is defined as the association strength of the hazard cause relative to the current hazard entity. In specific implementation, if multiple paths exist, the maximum value among the path weights or their weighted average is selected as the association strength. Optionally, the association strengths of all hazard causes associated with the hazard entity are weighted and fused to obtain the entity risk value of the hazard entity. It can be understood that the weights for weighted fusion can be set according to the type of hazard cause or historical statistical data. When calculating the entity risk value of the hazard entity, the weighted summation formula is used as follows: ; in: This represents the risk value of the i-th potential hazard entity. This represents the correlation strength weight of the j-th hazard cause. This indicates the correlation strength of the cause of the j-th hidden danger.

[0027] The normalization process uses the min-max standardization method, which is the maximum sum of historical entity risk values. Minimum value The total risk value of current image entities At that time, the probability of hidden dangers When fusing non-image dimensional information, the auxiliary decision node uses a linear regression model to process the quantized values. For example, the quantized value of ambient temperature information is... (Unit: °C) When The output adjustment factor is 1.2. The output coefficient is 1.0. The output coefficient is 0.8. The effective weight of the directed connection from the attribute feature node to the associated node is adjusted. For example, if the original weight is 0.7, the effective weight at a temperature of 65℃ is 0.7 × 1.2 = 0.84.

[0028] The sum of the entity risk values ​​of all identified hazard entities in the hazard image is calculated, and the sum is normalized and mapped to a value between zero and one hundred. This value is used as the hazard probability of the corresponding area in the hazard image. In some embodiments, the normalization process is based on a linear mapping between the maximum and minimum sum of entity risk values ​​in historical data, expressed by the following formula: ; in: This represents the calculated probability value of the potential hazard. This represents the sum of entity risk values ​​in the current hazard image. This represents the maximum sum of historical entity risk values. This represents the minimum sum of entity risk values ​​based on historical statistics.

[0029] After acquiring images of potential hazards in the target area, the process also includes acquiring non-image dimension information at the same spatiotemporal point as the hazard images. In specific implementations, non-image dimension information includes at least one of ambient temperature, equipment vibration, and gas concentration information. It can be understood that non-image dimension information is acquired synchronously through corresponding sensors deployed along the inspection path. In the hazard knowledge network, an auxiliary decision-making node is established for each type of non-image dimension information. The auxiliary decision-making node is connected to relevant attribute feature nodes and associated nodes. In some embodiments, the auxiliary decision-making node receives quantized non-image dimension information values ​​as input. When performing network inference to obtain the hazard type and hazard cause associated with the hazard entity, the quantized value of the non-image dimension information is used as the input to the corresponding auxiliary decision-making node. The auxiliary decision-making node outputs an adjustment coefficient based on the input value. This adjustment coefficient is used to dynamically adjust the effective weight of the directed connections from relevant attribute feature nodes to associated nodes in the current inference process. In specific implementations, the effective weight is the product of the original directed connection weight and the adjustment coefficient.

[0030] See Figure 4This is a comparison chart showing the adjustment of connection weights in a knowledge network within an intelligent hazard identification system. It illustrates the dynamic optimization of network connection weights by non-image-level information and is a core analytical chart in the hazard knowledge network inference stage. Gray represents the original weights, and blue represents the effective weights after adjustment by non-image information. Increased temperature sensor data strengthens the connection weight. Temperature sensor data is associated with the cause of "excessive temperature," increasing the weight. Stable gas sensor data decreases the connection weight. This chart demonstrates the dynamic optimization effect of non-image-level information on the hazard knowledge network: by adjusting the effective weights of network connections using real-time data from auxiliary decision nodes (such as sensors), the network inference results are made more closely aligned with the actual environment, improving the accuracy of hazard identification.

[0031] Example 4: In specific implementation, when the probability of a potential hazard exceeds a preset threshold, a task generation command is automatically triggered. In this implementation, the preset threshold is set based on historical investigation records and risk level classification. The triggering condition for the task generation command can be expressed by the formula: ; in: This indicates that the trigger status of the task generation command is 1 for triggered and 0 for not triggered. This represents the calculated probability value of the potential hazard. This represents a system-preset threshold. In response to a task generation command, the system integrates the hazard type, hazard cause, and related non-image dimension information associated with the hazard entities identified in the corresponding area of ​​the hazard image, forming a structured task description. This structured task description organizes information using predefined fields and formats. In some embodiments, based on the location information carried by the hazard image and the generated structured description, an intelligent investigation task is generated, containing the target location, target object, and key points of concern. The target location is obtained from the spatiotemporal metadata of the hazard image, the target object is the identified hazard entity type, and the key points of concern are the associated hazard causes and non-image dimension anomalies.

[0032] After outputting the intelligent investigation task, during the on-site investigation of the reported hazard entity, the on-site handling measures and the status after handling are recorded. In specific implementation, on-site handling measures include repair, replacement, isolation, and cleaning, and the status after handling is recorded through review images or text descriptions. Optionally, the recorded on-site handling measures are compared with the hazard causes associated with the hazard entity in the hazard knowledge network to mark the root cause node directly targeted by the on-site handling measures. In some embodiments, the root cause node is the associated node that the on-site handling measures aim to eliminate or control. The root cause node is associated with the corresponding hazard entity to form a verified knowledge association pair. The verified knowledge association pair includes the attribute characteristics of the hazard entity, the root cause node, and the handling effectiveness marker. In specific implementation, referring to Table 1, the verified knowledge association pair can be recorded as a data record containing specific fields.

[0033] Table 1: Verified Knowledge Association Pair Record Table ; The process of forming verified knowledge association pairs between root cause nodes and corresponding hazard entities involves storing the structured information shown in the table above as a complete record in a knowledge verification database. This database is used for subsequent optimization of the hazard knowledge network. In some embodiments, when the same hazard entity is associated with multiple hazard causes, multiple verified knowledge association pairs are generated. Each knowledge association pair records the correspondence between a specific root cause node and a corresponding remedial measure.

[0034] Example 5: In specific implementation, the verified knowledge association pairs are received from the field verification stage. The verified knowledge association pairs include root cause nodes and corresponding hazard entity identifiers and matching attribute features. In specific implementation, the verified knowledge association pair records are read from the knowledge verification database. The existing directed connections between the root cause nodes and attribute feature nodes of the verified knowledge association pair records are identified in the hazard knowledge network. It can be understood that the identification process searches in the topology of the hazard knowledge network based on the node name or unique identifier.

[0035] Based on the effectiveness verification results of the on-site treatment measures, the weights of the original directed connections are strengthened or weakened. When adjusting the weights of the original directed connections, the learning rate coefficient is... The reward factor is set to 0.1 based on the validity verification results. When the treatment is effective When the treatment is ineffective Penalty coefficient The average weight of the network is obtained by calculating the arithmetic mean of the weights of all directed connections. The statistical period for redundant connections is set to 30 days, the weight decay factor is set to 0.8, and the removal threshold is set to 0.05. For example, if a directed connection initially has a weight of 0.3 and is not associated for three consecutive statistical periods, the weight decays to 0.24, 0.192, and 0.1536 respectively. If it is still higher than the threshold of 0.05, it is retained. If it decays to 0.0983 and 0.0786 for another two periods and is still not associated, it continues to decay until it falls below 0.05, at which point it is removed. After each weight adjustment, L2 regularization is used to constrain the weights of all connections in the entire hidden danger knowledge network to prevent excessively large weights of individual connections from causing bias in network inference. The parameter settings are determined based on the association patterns of historical hidden danger cases and field verification feedback data, combined with the stability requirements of network inference. Weight adjustments are based on the effectiveness verification results; effective weights are strengthened, and ineffective weights are weakened. Redundant connections decay gradually at fixed intervals. In some embodiments, the validity verification result is recorded in the validity tag field of the verified knowledge association pair. If the validity tag confirms that the action measure is effective, the weight of the corresponding original directed connection is increased; if the validity tag confirms that the action measure is ineffective, the weight of the corresponding original directed connection is decreased. Optionally, the weight adjustment is calculated according to an incremental formula, expressed as: ; in: This represents the adjusted directed connection weights. This represents the original directed connection weights before adjustment. This represents the learning rate coefficient. This represents the reward factor based on the validity verification results. Indicates the penalty coefficient. This represents the absolute value of the deviation between the current directed connection weights and the network average weights. In practice, each time a new validity verification result is received, the relevant original directed connection weights are updated and recalculated.

[0036] For redundant connections that exist in the hazard knowledge network but have never been verified, their weights are gradually reduced until they are removed. In some embodiments, the system periodically scans all directed connections in the hazard knowledge network for usage records. A usage record refers to the number of times a directed connection is activated and subsequently verified during network inference. Optionally, for directed connections with zero usage records within a statistical period, their weights are gradually reduced by multiplying them by a weight decay factor less than 1. When the weight of a directed connection falls below a set removal threshold after continuous decay, the directed connection is removed from the hazard knowledge network.

[0037] See Figure 5This is a bar chart showing the connection weight distribution of the intelligent hazard investigation knowledge network. It displays the weight values ​​of each connection in the hazard knowledge network and is a core analytical chart in the knowledge network construction phase. The red dashed line represents the average weight of the network, used to compare the weight levels of each connection. The highest weight indicates the strongest correlation between the auxiliary node and the feature. The gas leak feature has a strong correlation with the cause of pressure anomalies. The lowest weight indicates a weak correlation between equipment cracks and improper installation. This chart visually presents the initial weight distribution of the hazard knowledge network: connections with higher weights represent a closer correlation between the corresponding feature and the cause / auxiliary information, which is the core basis for subsequent network reasoning (identifying hazard types and causes). At the same time, the average weight line can help identify "strong correlation" and "weak correlation" connections, providing a benchmark for subsequent weight optimization.

[0038] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0039] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent hazard identification based on image recognition, characterized in that, Includes the following steps: Establish a hazard knowledge network to represent the relationship between hazard types, hazard characteristics, and hazard causes; Images of potential hazards in the target area are collected along the designated inspection route, and these images are then input into the hazard knowledge network. From the initial analysis results output by the hazard knowledge network, extract the hazard entities present in the hazard image; The entity features of the hidden danger entity are matched with the hidden danger features in the hidden danger knowledge network, and reasoning is performed through the hidden danger knowledge network to obtain the hidden danger type and hidden danger cause associated with the hidden danger entity. Based on the correlation strength of the causes of the hidden dangers, the probability of hidden dangers in the corresponding area of ​​the hidden danger image is calculated; Based on the comparison between the probability of the potential hazard and a preset threshold, an intelligent investigation task is generated that is bound to the location of the potential hazard image. The intelligent investigation task is output to guide the targeted investigation of potential hazards in the target area; Based on the on-site verification feedback for the intelligent investigation task, the relationships in the hidden danger knowledge network are optimized.

2. The intelligent hazard investigation method based on image recognition according to claim 1, characterized in that, The establishment of a hazard knowledge network to represent the relationship between hazard types, hazard characteristics, and hazard causes includes: Obtain historical hazard cases, and analyze the types of historical hazards, historical image features, and historical causes from these cases; The historical hazard type is defined as a node type in the hazard knowledge network, the historical image features are abstracted as attribute features in the hazard knowledge network, and the historical causes are established as associated nodes in the hazard knowledge network. A directed connection is established between the node type, the attribute feature, and the associated node. The weight of the directed connection is initialized based on the co-occurrence frequency of the corresponding item in the historical hidden danger cases.

3. The intelligent hazard investigation method based on image recognition according to claim 2, characterized in that, Extracting the hazard entities present in the hazard image from the initial analysis results output from the hazard knowledge network includes: Receive the initial analysis results of the hidden danger knowledge network on the hidden danger image, the initial analysis results including multiple image regions in the hidden danger image that have a potential matching degree with the attribute features; Entity boundaries are defined for each of the multiple image regions, and local feature enhancement is performed within the regions defined by the entity boundaries. Identify whether there is an entity feature that matches any of the attribute features within the region after the local feature enhancement. If the match is successful, define the region bounded by the entity boundary as a potential entity.

4. The intelligent hazard investigation method based on image recognition according to claim 3, characterized in that, The step of matching the entity features of the potential hazard with the potential hazard features in the potential hazard knowledge network, and then using the potential hazard knowledge network to infer the potential hazard type and cause associated with the potential hazard, includes: The similarity between the entity features and all attribute features in the hidden danger knowledge network is calculated, and the attribute feature with the highest similarity is selected as the matching result. Locate the attribute feature node corresponding to the matching result in the hidden danger knowledge network, and activate all directed connections originating from the attribute feature node; Traverse the path along the activated directed connection, gather all reachable node types and associated nodes, take the gathered node types as the hazard types associated with the hazard entity, and take the gathered associated nodes as the hazard causes associated with the hazard entity.

5. The intelligent hazard investigation method based on image recognition according to claim 4, characterized in that, The step of calculating the probability of a hazard in the region corresponding to the hazard image based on the correlation strength of the hazard causes includes: For each hazard cause associated with the hazard entity, the path weight from the attribute feature node corresponding to the matching result to the associated node corresponding to the hazard cause is obtained in the hazard knowledge network, and the path weight is defined as the association strength of the hazard cause. The entity risk value of the entity is obtained by weighted fusion of the correlation strength of all the causes of the hidden danger associated with the entity. The sum of the entity risk values ​​of all the hazard entities in the hazard image is calculated, and the sum is normalized and mapped to a value between zero and one hundred. This value is then used as the hazard probability of the corresponding area in the hazard image.

6. The intelligent hazard investigation method based on image recognition according to claim 5, characterized in that, After acquiring images of potential hazards in the target area, the method further includes: Collect non-image dimension information at the same spatiotemporal point as the hazard image, wherein the non-image dimension information includes at least one of ambient temperature information, equipment vibration information, and gas concentration information; In the aforementioned hidden danger knowledge network, auxiliary decision-making nodes are established for each type of non-image dimension information; During network inference, the quantized value of the non-image dimension information is used as the input of the auxiliary decision node to adjust the effective weight of the directed connection from the attribute feature node to the associated node.

7. The intelligent hazard investigation method based on image recognition according to claim 6, characterized in that, The step of generating an intelligent investigation task bound to the location of the hazard image based on the comparison result between the numerical value of the hazard probability and a preset threshold includes: When the probability of the potential hazard is greater than a preset threshold, a task generation instruction is triggered; In response to the task generation instruction, the hazard type, the cause of the hazard, and the non-image dimension information associated with the hazard entity in the region corresponding to the hazard image are integrated to form a structured description; Based on the location information of the hazard image and the structured description, an intelligent investigation task is generated that includes the target location, target object, and key points of concern.

8. The intelligent hazard investigation method based on image recognition according to claim 7, characterized in that, After outputting the intelligent screening task, the method further includes: During the on-site investigation of the aforementioned intelligent investigation task, the on-site handling measures and post-handling status for the potential hazards are recorded. The on-site handling measures are compared with the hazard causes associated with the hazard knowledge network, and the root cause nodes targeted by the on-site handling measures are marked. The root cause node is associated with the corresponding hidden danger entity to form a verified knowledge association pair.

9. The intelligent hazard investigation method based on image recognition according to claim 8, characterized in that, The optimization of the relationships in the hazard knowledge network based on on-site verification feedback for the intelligent investigation task includes: Receive the verified knowledge association pair and identify the original directed connection between the root cause node and the attribute feature node; Based on the effectiveness verification results of the aforementioned on-site handling measures, the weights of the original directed connections may be increased or decreased. For redundant connections that exist in the hidden danger knowledge network but have never been associated with the verified knowledge association pairs, their weights are gradually reduced until they are removed.

10. An intelligent hazard detection system based on image recognition, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent hazard investigation method based on image recognition as described in any one of claims 1 to 9.

Images