Intelligent dynamic pilot frequency networking system and method for MESH ad hoc network of image transmission module

Abstract

The invention relates to the technical field of unmanned aerial vehicles, in particular to an intelligent dynamic pilot frequency networking system and method for an image transmission module MESH ad hoc network. The method is based on spectrum fingerprint features, combines a real-time interference topological graph constructed by interaction of neighborhood nodes, identifies a conflict frequency set needing to be evaded through multi-dimensional weighted evaluation, and dynamically generates a pilot frequency strategy table containing main and standby frequencies and a switching threshold value by using a frequency hopping decision engine according to the conflict frequency set and a preset legal frequency pool. Synchronizing the pilot frequency strategy table to an associated relay node, generating a link stability evaluation matrix and feeding back the link stability evaluation matrix to a frequency hopping decision engine; and correcting a frequency switching threshold value and an alternative frequency weight in the pilot frequency strategy table based on the link stability evaluation matrix. According to the invention, the spectrum utilization efficiency of the unmanned aerial vehicle image transmission system can be improved, the stability and reliability of a multi-hop link are ensured, the networking time delay is reduced through a distributed decision-making mechanism, and an effective solution is provided for high-quality wireless transmission in a dynamic topology environment.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles (UAVs), and in particular to an intelligent dynamic heterogeneous frequency networking system and method for a MESH self-organizing network of image transmission modules. Background Art

[0002] With the rapid development of fields such as drone swarms, emergency communications, and the Industrial Internet of Things, mesh self-organizing networking technology has become a key solution for wireless communications due to its decentralization, high reliability, and flexible scalability. However, in complex electromagnetic environments, traditional mesh networks often face challenges such as spectrum interference, channel congestion, and dynamic topology changes, leading to degraded communication quality and even link interruptions. Existing technologies use fixed frequency bands or simple frequency hopping mechanisms that struggle to adapt to ever-changing interference environments, while static spectrum allocation schemes cannot meet the low latency and high stability requirements of highly dynamic scenarios. Furthermore, traditional mesh networks often lack efficient management of multiple nodes and hops, limiting network scalability and making it difficult to support large-scale drone collaborative operations or long-distance image transmission communications.

[0003] To address these issues, existing technologies have attempted to employ dynamic spectrum sensing or interference avoidance algorithms. However, these approaches often rely on centralized control, which carries the risk of decision delays and single-point failures. Distributed solutions, on the other hand, often lack the ability to perform global interference collaborative analysis, resulting in local optimization but poor overall performance. Furthermore, existing image transmission modules still lack cross-layer link quality assessment, intelligent frequency switching, and adaptive topology management, making it difficult to achieve stable and efficient data transmission in complex environments. Therefore, there is an urgent need for an intelligent dynamic inter-frequency networking system and method to address these pain points and provide high-performance wireless networking support for scenarios such as drone swarms and emergency communications. Summary of the Invention

[0004] The present invention overcomes the deficiencies of the prior art and provides a system and method for intelligent dynamic heterodyne frequency networking of a MESH self-organizing network of image transmission modules.

[0005] In order to achieve the above-mentioned purpose, the technical solution adopted by the present invention is: The first aspect of the present invention discloses a method for intelligent dynamic inter-frequency networking of a MESH self-organizing network of an image transmission module, comprising the following steps: S102: Scanning the original time domain signal data of the preset frequency band of each target node in real time to generate spectrum fingerprint features including interference intensity distribution and channel occupancy pattern; S104: Based on the spectrum fingerprint features and in combination with the real-time interference topology map constructed through interaction between neighboring nodes, a conflicting frequency set to be avoided is identified through multi-dimensional weighted evaluation; S106: Dynamically generate an inter-frequency strategy table including primary and backup frequencies and switching thresholds using a frequency hopping decision engine based on the conflicting frequency set and a preset legal frequency pool, and synchronize the inter-frequency strategy table to associated relay nodes; S108: Collecting physical layer bit error rate and network layer delay data through a cross-layer frame detection mechanism, generating a link stability evaluation matrix and feeding it back to the frequency hopping decision engine; S110: Dynamically modifying the frequency switching threshold and the alternative frequency weight in the inter-frequency strategy table based on the historical sequence of the link stability evaluation matrix and the environmental change parameters.

[0006] Preferably, the S102 is specifically: Each target node uses a distributed RF sampling unit to synchronously sample the preset frequency band within a preset time window to obtain the original time domain signal data; Perform fast Fourier transform on the original time domain signal data to obtain the instantaneous signal energy value of each discrete frequency point and generate a spectrum energy distribution diagram; Based on the spectrum energy distribution diagram, the average background noise energy in the noise assessment sub-band is counted as the dynamic noise floor, and the dynamic interference determination threshold is set according to the noise floor; Scan the entire frequency band, identify areas where energy continuously exceeds the dynamic interference determination threshold as interference pulses, and record the center frequency, bandwidth, and peak energy of each interference pulse; For each interference pulse, the relative intensity ratio of its peak energy to the dynamic noise floor is calculated as the interference intensity value. The center frequency, bandwidth, and interference intensity values ​​of all interference pulses are combined and mapped to the frequency domain coordinates to generate the interference intensity distribution characteristics. The spectrum energy distribution diagram is segmented according to preset channel rules. The energy value of each channel is continuously monitored. If there is an interference pulse or the average energy exceeds the channel busy-idle threshold (the busy-idle threshold is set based on the dynamic noise floor), the channel is marked as "busy"; otherwise, it is marked as "idle". The start and end times of the state changes are recorded to form an occupancy event list. The occupancy rate, average occupancy time, and idle-busy transition frequency of each channel are calculated to generate a channel occupancy pattern. The interference intensity distribution characteristics are associated with the channel occupancy pattern and fused into a structured data object, which is output as the spectrum fingerprint feature.

[0007] Preferably, the S104 is specifically: Extract the interference intensity distribution characteristics from the spectrum fingerprint characteristics of the local node, analyze the center frequency, bandwidth and interference intensity value of each interference pulse, and generate a local interference source feature table. At the same time, receive the interference source feature table broadcast by the neighboring node through the dedicated control channel to form a set of neighboring interference sources. Determine the signal attenuation factor of the neighboring interference source on the current node based on the geographical location relationship between the nodes, and correct the interference intensity value in the neighboring interference source characteristic table based on the attenuation factor to generate an equivalent interference intensity value; integrate the corrected neighboring interference source and the current node interference source according to the frequency domain coordinates, and map them into a real-time interference topology map with frequency as the horizontal axis and equivalent interference intensity as the vertical axis; The equivalent interference intensity value corresponding to each frequency point in the real-time interference topology map is read as the basic conflict weight; the channel occupancy pattern in the spectrum fingerprint feature is associated to extract the occupancy rate and idle-busy transition frequency of the channel to which the corresponding frequency belongs, thereby determining the channel instability coefficient; at the same time, the number of times the corresponding frequency point appears in the neighborhood interference source feature table is counted to generate the neighborhood conflict density value; The basic conflict weight, channel instability coefficient and neighborhood conflict density value of each frequency point are weighted to obtain the spectrum conflict risk value of each frequency point; the frequency points whose spectrum conflict risk value exceeds the preset risk threshold are screened, and the channel bandwidths to which they belong are merged to form a set of conflict frequencies that need to be avoided.

[0008] Preferably, the step S106 is specifically: Eliminate all frequency points and associated channels covered by the conflicting frequency set from a preset legal frequency pool to generate a candidate frequency set; For each frequency point in the candidate frequency set, calculate its minimum frequency domain separation distance from the conflicting frequency set, and combine the average bit error rate performance of this frequency point in the historical link stability evaluation matrix to generate the frequency availability coefficient; The frequency availability coefficients of each frequency point in the candidate frequency set are sorted in descending order, and the frequency point with the highest ranking is selected as the primary frequency. Among the remaining frequency points, based on the preset frequency band dispersion constraint rule (which requires the candidate frequencies to maintain a minimum frequency band spacing with the primary frequency and with each other), the two frequency points with the highest frequency availability coefficients that meet the constraint are selected as the primary and secondary backup frequencies. The real-time physical layer bit error rate baseline value collected by the associated cross-layer frame detection mechanism is multiplied by the preset degradation margin factor by the historical average bit error rate corresponding to the primary frequency to generate the primary frequency switching trigger threshold; based on the predicted fluctuation range of the link quality of the first-level backup frequency, the switching hysteresis interval between it and the primary frequency is set (to prevent frequent switching).

[0009] The main frequency, two-level backup frequencies, main frequency switching trigger threshold and switching hysteresis interval are integrated into a structured inter-frequency strategy table; the inter-frequency strategy table is encrypted and encapsulated into a strategy update instruction frame through a dedicated control channel and broadcast to the associated relay nodes in real time.

[0010] Preferably, the step S108 is specifically: A dedicated probe frame carrying a unique sequence number and dual timestamps is constructed at the media access control layer and injected into the transmit queue of the link to be evaluated. The dual timestamps include a transmit timestamp T1 and a scheduling timestamp T2. The scheduling timestamp T2 records the system time when the frame enters the physical layer buffer. When the receiving end's physical layer parses the arriving probe frame, it counts the original number of errored bits of the frame based on the forward error correction decoding result, and calculates the instantaneous bit error rate based on the probe frame length; it synchronously records the actual reception completion timestamp T3 of the frame, and extracts the sequence number to generate the original errored sample; The network layer captures the reception event of the probe frame and matches T1 and T2 registered by the sender according to the sequence number; calculates the difference between T3 and T1 to generate the end-to-end transmission delay, and calculates the difference between T3 and T2 to generate the air interface delay, forming a delay vector; Using the sequence number as a key, the original bit error samples corresponding to the same probe frame are bound to the delay vector to generate a link quality tuple including the bit error rate, transmission delay, and air delay; Within a preset evaluation window, link quality tuples of multiple links are aggregated. For each link, the coefficient of variation of the bit error rate is calculated as the bit error fluctuation factor, the percentile range of the transmission delay is calculated as the delay jitter coefficient, and the number of times the air delay exceeds the threshold is counted to generate the channel contention index. A two-dimensional matrix is ​​constructed with link IDs as rows and stability indicators as columns. Each row is filled with the bit error fluctuation factor, delay jitter coefficient, and channel contention index of the corresponding link. The output is the link stability evaluation matrix and input into the frequency hopping decision engine.

[0011] Preferably, the S110 is specifically: Extracting the historical bit error fluctuation factor and delay jitter coefficient of each link in the link stability evaluation matrix, calculating the time-attenuated weighted average of the historical bit error fluctuation factor and delay jitter coefficient, and generating a stability degradation gradient that reflects the long-term evolution trend of link quality; Aligning the interference intensity distribution feature in the current spectrum fingerprint feature with the stability degradation gradient in the time domain, establishing a mapping relationship between interference pulse energy and bit error fluctuation factor, and generating a frequency band vulnerability coefficient that characterizes the interference sensitivity of a specific frequency band; For the main frequency switching trigger threshold in the inter-frequency strategy table, the historical deterioration rate of the main frequency switching trigger threshold is determined based on the stability degradation gradient. Combined with the real-time error fluctuation factor of the current link, the deterioration margin factor is dynamically corrected through the proportional-integral regulator to output a dynamic switching threshold that is adaptive to changes in link quality; Based on the frequency band vulnerability coefficient, the frequency points that maintain the minimum protection interval with the frequency band where the interference pulse whose interference intensity value exceeds the preset interference threshold is located are screened out from the candidate frequency set, and the delay jitter coefficient corresponding to the link quality history data of the frequency point is extracted and the inverse is taken to generate the delay stability weight; at the same time, the frequency of occurrence of the frequency point in the heterodyne strategy table of the neighboring node is counted to generate the topology compatibility weight; The delay stability weight and the topology compatibility weight are linearly combined and normalized to obtain the comprehensive priority score of each alternative frequency. The fixed threshold in the original inter-frequency strategy table is replaced by a dynamic switching threshold, and the alternative frequency list is reordered according to the comprehensive priority score. The two highest-scoring backup frequencies are retained and their switching hysteresis intervals are updated to form an optimized inter-frequency strategy table.

[0012] A second aspect of the present invention discloses an intelligent dynamic inter-frequency networking system for a MESH self-organizing network of an image transmission module, which is applied to any one of the above-mentioned methods for intelligent dynamic inter-frequency networking of a MESH self-organizing network of an image transmission module. The system includes: Distributed RF sampling unit, integrated with a wideband tunable RF front end, supports simultaneous real-time spectrum scanning within a preset frequency band, acquires raw time-domain signal data, and generates spectrum energy distribution maps based on fast Fourier transform; A spectrum fingerprint analysis module, connected to the distributed RF sampling unit, is used to calculate the dynamic noise floor, interference pulse characteristics and channel occupancy mode, and output structured spectrum fingerprint characteristics; The neighborhood interference topology construction module receives the interference source feature table of the neighboring nodes through a dedicated control channel, corrects the interference intensity value based on the signal attenuation factor, and generates a real-time interference topology map; A frequency hopping decision engine, connected to the conflicting frequency evaluation module, is used to screen a candidate frequency set from a preset legal frequency pool, calculate a frequency availability coefficient, and dynamically generate an inter-frequency strategy table including a primary and backup frequency and a switching threshold; Network topology management unit, supporting 2 to 264 nodes; 1 hop, 2 hops, 3 hops, 4 hops or unlimited hops; and supports software distance limit, up to 500KM; Multi-protocol compatible interface supports hardware adaptation and software protocol stack expansion with image transmission modules.

[0013] The present invention solves the technical deficiencies in the background technology and has the following beneficial effects: real-time scanning of the original time-domain signal data of the preset frequency band of each target node to generate spectrum fingerprint features including interference intensity distribution and channel occupancy pattern; based on the spectrum fingerprint features, combined with the real-time interference topology map constructed by the interaction of neighboring nodes, a multi-dimensional weighted evaluation is performed to identify the conflicting frequency set to be avoided; based on the conflicting frequency set and the preset legal frequency pool, a frequency hopping decision engine is used to dynamically generate an inter-frequency strategy table including the primary and backup frequencies and switching thresholds, and the inter-frequency strategy table is synchronized to the associated relay nodes; a cross-layer frame detection mechanism is used to collect physical layer bit error rate and network layer delay data to generate a link stability evaluation matrix that is fed back to the frequency hopping decision engine; based on the historical sequence of the link stability evaluation matrix and environmental change parameters, the frequency switching threshold and alternative frequency weight in the inter-frequency strategy table are dynamically corrected. The present invention can improve the spectrum utilization efficiency of the drone image transmission system, ensure the stability and reliability of the multi-hop link, and reduce the networking delay through a distributed decision mechanism, providing an effective solution for high-quality wireless transmission in a dynamic topology environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, without paying any creative work, they can also obtain drawings of other embodiments based on these drawings.

[0015] Figure 1 This is the overall method flow chart of the MESH self-organizing network intelligent dynamic frequency networking method of this image transmission module; Figure 2 This is a partial flow chart of the method for the MESH self-organizing network intelligent dynamic frequency networking method of this image transmission module; Figure 3 This is the system block diagram of the MESH self-organizing network intelligent dynamic heterogeneous frequency networking system of this image transmission module. DETAILED DESCRIPTION

[0016] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features therein can be combined with each other.

[0017] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0018] like Figure 1 As shown, the first aspect of the present invention discloses a method for intelligent dynamic heterodyne frequency networking of a MESH self-organizing network of an image transmission module, comprising the following steps: S102: Scanning the original time domain signal data of the preset frequency band of each target node in real time to generate spectrum fingerprint features including interference intensity distribution and channel occupancy pattern; S104: Based on the spectrum fingerprint features and in combination with the real-time interference topology map constructed through interaction between neighboring nodes, a conflicting frequency set to be avoided is identified through multi-dimensional weighted evaluation; S106: Dynamically generate an inter-frequency strategy table including primary and backup frequencies and switching thresholds using a frequency hopping decision engine based on the conflicting frequency set and a preset legal frequency pool, and synchronize the inter-frequency strategy table to associated relay nodes; S108: Collecting physical layer bit error rate and network layer delay data through a cross-layer frame detection mechanism, generating a link stability evaluation matrix and feeding it back to the frequency hopping decision engine; S110: Dynamically modifying the frequency switching threshold and the alternative frequency weight in the inter-frequency strategy table based on the historical sequence of the link stability evaluation matrix and the environmental change parameters.

[0019] Preferably, the step S102 is specifically: Each target node uses a distributed RF sampling unit to synchronously sample the preset frequency band within a preset time window to obtain the original time domain signal data; The target node refers to each drone image transmission module terminal participating in the dynamic networking in the Mesh network; Perform fast Fourier transform on the original time domain signal data to obtain the instantaneous signal energy value of each discrete frequency point and generate a spectrum energy distribution diagram; Here, discrete frequency points refer to equally spaced frequency analysis units obtained by Fourier transforming the sampling frequency band; Based on the spectrum energy distribution diagram, the average background noise energy in the noise assessment sub-band is counted as the dynamic noise floor, and the dynamic interference determination threshold is set according to the noise floor; Scan the entire frequency band, identify areas where energy continuously exceeds the dynamic interference determination threshold as interference pulses, and record the center frequency, bandwidth, and peak energy of each interference pulse; For each interference pulse, the relative intensity ratio of its peak energy to the dynamic noise floor is calculated as the interference intensity value. The center frequency, bandwidth, and interference intensity values ​​of all interference pulses are combined and mapped to the frequency domain coordinates to generate the interference intensity distribution characteristics. The spectrum energy distribution diagram is segmented according to preset channel rules. The energy value of each channel is continuously monitored. If there is an interference pulse or the average energy exceeds the channel busy-idle threshold (the busy-idle threshold is set based on the dynamic noise floor), the channel is marked as "busy"; otherwise, it is marked as "idle". The start and end times of the state changes are recorded to form an occupancy event list. The occupancy rate, average occupancy time, and idle-busy transition frequency of each channel are calculated to generate a channel occupancy pattern. Among them, the preset channel rules refer to a fixed bandwidth channel division method pre-divided according to wireless communication standards or system requirements. For example, in the 5.8GHz frequency band, the 5.725-5.850GHz frequency band is divided into 6 independent channels according to 20MHz bandwidth.

[0020] The interference intensity distribution characteristics are associated with the channel occupancy pattern and fused into a structured data object, which is output as the spectrum fingerprint feature.

[0021] In one embodiment of the present invention, a mesh network of ten drones equipped with image transmission modules was deployed in a drone swarm urban inspection scenario. Each node synchronously performed wideband sampling within a 500ms time window using a distributed RF sampling unit (operating in the 5.725-5.850GHz frequency band) to acquire raw time-domain signal data. After generating a spectrum energy distribution map using a fast Fourier transform, the 5.740-5.745GHz sub-band was selected, and the average background noise energy was calculated to be -95dBm. Based on this, the dynamic interference threshold was set to -80dBm. A full-band scan revealed an interference pulse at 5.810GHz with a duration of 20ms, a bandwidth of 15MHz, and a peak energy of -65dBm (interference strength of 30dB). The spectrum was then segmented into 20MHz channels. The 5.810-5.830GHz channel was marked as "busy" due to the interference pulse and average energy exceeding the threshold (-78dBm). Statistics show that its occupancy rate reached 85%, with an idle-to-busy transition frequency of 12 times per second. Finally, the interference pulse characteristics (center frequency, bandwidth, and intensity) are fused with the channel occupancy pattern to generate structured spectrum fingerprint features.

[0022] In summary, this method can accurately identify interference sources in the environment and dynamically generate spectrum fingerprint features through distributed RF sampling and real-time spectrum analysis, achieving adaptive perception of complex electromagnetic environments, thereby providing a highly reliable interference avoidance basis for subsequent intelligent frequency switching.

[0023] Preferably, the S104 is specifically: Extract the interference intensity distribution characteristics from the spectrum fingerprint characteristics of the local node, analyze the center frequency, bandwidth and interference intensity value of each interference pulse, and generate a local interference source feature table. At the same time, receive the interference source feature table broadcast by the neighboring node through the dedicated control channel to form a set of neighboring interference sources. Among them, the local node refers to the Mesh network member device (such as drone No. 5 in the embodiment) that currently performs spectrum analysis and generates the interference signature table. It is an autonomous decision-making entity relative to other neighboring nodes within its communication range.

[0024] Determine the signal attenuation factor of the neighboring interference source on the current node based on the geographical location relationship between the nodes, and correct the interference intensity value in the neighboring interference source characteristic table based on the attenuation factor to generate an equivalent interference intensity value; integrate the corrected neighboring interference source and the current node interference source according to the frequency domain coordinates, and map them into a real-time interference topology map with frequency as the horizontal axis and equivalent interference intensity as the vertical axis; The equivalent interference intensity value corresponding to each frequency point in the real-time interference topology map is read as the basic conflict weight; the channel occupancy pattern in the spectrum fingerprint feature is associated to extract the occupancy rate and idle-busy transition frequency of the channel to which the corresponding frequency belongs, thereby determining the channel instability coefficient; at the same time, the number of times the corresponding frequency point appears in the neighborhood interference source feature table is counted to generate the neighborhood conflict density value; The basic conflict weight, channel instability coefficient and neighborhood conflict density value of each frequency point are weighted to obtain the spectrum conflict risk value of each frequency point; the frequency points whose spectrum conflict risk value exceeds the preset risk threshold are screened, and the channel bandwidths to which they belong are merged to form a set of conflict frequencies that need to be avoided.

[0025] In one embodiment of the present invention, in a drone swarm urban inspection scenario, drone node 5 analyzes its spectrum fingerprint to identify local interference sources (center frequency 5.810 GHz, bandwidth 15 MHz, interference strength 30 dB). It also receives interference source signature tables broadcast by neighboring nodes 2, 7, and 9 via a dedicated control channel (including interference data for 5.805 GHz / 20 MHz / 28 dB and 5.815 GHz / 10 MHz / 32 dB, respectively). Based on GPS positioning, it calculates the distance between the neighboring nodes and the drone itself (using an attenuation factor of 0.8 within 300 meters), corrects the neighboring interference strength to an equivalent value (e.g., 32 dB for node 7 is corrected to 25.6 dB), and integrates these values ​​to generate a real-time interference topology map. Correlating the channel occupancy pattern reveals an 85% occupancy rate for the 5.810-5.830 GHz channel and an idle-to-busy transition frequency of 12 times per second (channel instability coefficient 0.92). Furthermore, the frequency appears three times in the neighboring interference table (neighborhood conflict density 0.75). After weighted calculation, the spectrum conflict risk value is 0.87 (exceeding the threshold of 0.7). Ultimately, the 20MHz channel to which this frequency band belongs is included in the conflict frequency set to avoid it.

[0026] In general, this method integrates the interference characteristics of local and neighboring nodes and takes geographical location factors into consideration to construct a real-time interference topology map. It then performs a multi-dimensional weighted assessment based on channel occupancy and neighborhood conflict density. This method can accurately identify high-conflict risk frequency bands and generate a set of conflict frequencies that need to be avoided, thereby improving the system's spectrum decision-making accuracy and anti-interference capability in complex interference environments.

[0027] Preferably, the S106, as Figure 2 As shown, specifically: S202: Eliminate all frequency points and associated channels covered by the conflicting frequency set from a preset legal frequency pool to generate a candidate frequency set; wherein the preset legal frequency pool refers to a pre-configured set of available frequency bands that comply with radio management regulations (e.g., six 20 MHz channels divided between 5.725-5.850 GHz); S204: For each frequency point in the candidate frequency set, calculate its minimum frequency domain separation distance from the conflicting frequency set, and generate a frequency availability coefficient based on the average bit error rate performance of the frequency point in the historical link stability evaluation matrix. The historical link stability evaluation matrix is ​​a time series data set that stores stability indicators such as bit error rate and delay of each link in a historical period, and is used to quantify the long-term communication quality of the frequency point. S206: Sort the frequency availability coefficients of the frequency points in the candidate frequency set in descending order, and select the frequency point with the highest frequency availability coefficient as the primary frequency. Among the remaining frequency points, based on a preset frequency band dispersion constraint rule, select the two frequency points that meet the constraint and have the highest frequency availability coefficients as the primary and secondary backup frequencies. The preset frequency band dispersion constraint rule requires a minimum frequency band separation (e.g., ≥40 MHz) between the primary and backup frequencies to ensure spectrum dispersion and avoid common-mode interference. S208: Associate the real-time physical layer bit error rate reference value collected by the cross-layer frame detection mechanism, multiply the historical average bit error rate corresponding to the primary frequency by the preset degradation margin factor to generate the primary frequency switching trigger threshold; based on the link quality prediction fluctuation range of the first-level backup frequency, set the switching hysteresis interval between it and the primary frequency (to prevent frequent switching); the real-time physical layer bit error rate reference value collected by the cross-layer frame detection mechanism refers to the current bit error rate (e.g., 1.5×10⁻) measured in real time at the physical layer through dedicated detection frames. 5 ), as the basis for calculating the switching threshold; S210: The primary frequency, two-level backup frequencies, primary frequency switching trigger threshold, and switching hysteresis interval are integrated into a structured inter-frequency strategy table. The inter-frequency strategy table is encrypted and encapsulated into a strategy update command frame via a dedicated control channel and broadcast in real time to associated relay nodes. Associated relay nodes are adjacent Mesh devices that have a direct communication link with the local node and require strategy synchronization (e.g., drones 2, 7, and 9 of drone 5 in this example).

[0028] Taking the aforementioned drone swarm urban inspection scenario as an example, drone node 5 removes CH4 (5.810-5.830 GHz), which is covered by the conflicting frequency set, from the preset legal frequency pool (six 20 MHz channels divided in the 5.725-5.850 GHz frequency band), and generates a candidate frequency set (CH1 / CH2 / CH3 / CH5 / CH6). For CH1 (5.725-5.745 GHz), the minimum interval between it and the conflicting frequency band is calculated to be 65 MHz. Combined with the average bit error rate of 1.2×10⁻ in the historical link stability matrix, 5 , generating a frequency availability coefficient of 0.92 (the highest); CH1 was selected as the primary frequency. Among the remaining channels, based on the frequency band dispersion constraint (minimum interval between primary and backup frequencies ≥ 40 MHz), CH5 (5.845-5.865 GHz, availability coefficient 0.85) and CH6 (5.865-5.885 GHz, coefficient 0.81) were selected as the primary and secondary backup frequencies, respectively. Based on the real-time bit error rate benchmark value of 1.5×10⁻ 5 , the historical average bit error rate of the main frequency (1.0×10⁻ 5 ) multiplied by the degradation margin factor 1.8, yielding a switching threshold of 1.8×10⁻ 5 ; Set the CH5 switching hysteresis interval to [1.5×10⁻ 5 ,2.0×10⁻ 5 Finally, the inter-frequency strategy table is generated through integration, encrypted with AES-256, and broadcast to the associated relay nodes 2, 7, and 9 through a dedicated control channel.

[0029] It should be noted that this method generates the optimal primary and backup frequency combination and switching strategy by intelligently screening the candidate frequency set and comprehensively considering the frequency domain interval, historical link quality and real-time bit error rate performance. It can effectively avoid interference frequency bands and ensure the stability of communication links. At the same time, through the frequency band dispersion constraint and switching hysteresis interval setting, it not only ensures the reasonable distribution of spectrum resources, but also avoids unnecessary frequent switching, thereby improving the communication reliability and spectrum utilization efficiency of the Mesh network in a dynamic interference environment.

[0030] Preferably, the step S108 is specifically: A dedicated probe frame carrying a unique sequence number and dual timestamps is constructed at the media access control layer and injected into the transmit queue of the link to be evaluated. The dual timestamps include a transmit timestamp T1 and a scheduling timestamp T2. The scheduling timestamp T2 records the system time when the frame enters the physical layer buffer. When the receiving end's physical layer parses the incoming probe frame, it counts the original number of errored bits in the frame based on the forward error correction decoding result and calculates the instantaneous bit error rate based on the probe frame length. It also synchronously records the actual reception completion timestamp T3 of the frame and extracts the sequence number to generate the original errored sample. The forward error correction decoding result refers to the original number of errored bits and their location information obtained after the receiving end performs error detection and correction on the received data using a preset error correction coding program (such as LDPC or RS code). This is used to accurately calculate the actual error situation in wireless transmission. The network layer captures the reception event of the probe frame and matches T1 and T2 registered by the sender according to the sequence number; calculates the difference between T3 and T1 to generate the end-to-end transmission delay, and calculates the difference between T3 and T2 to generate the air interface delay, forming a delay vector; Using the sequence number as a key, the original bit error samples corresponding to the same probe frame are bound to the delay vector to generate a link quality tuple including the bit error rate, transmission delay, and air delay; Within a preset evaluation window, link quality tuples of multiple links are aggregated. For each link, the coefficient of variation of the bit error rate is calculated as the bit error fluctuation factor, the percentile range of the transmission delay is calculated as the delay jitter coefficient, and the number of times the air delay exceeds the threshold is counted to generate the channel contention index. A two-dimensional matrix is ​​constructed with link IDs as rows and stability indicators as columns. Each row is filled with the bit error fluctuation factor, delay jitter coefficient, and channel contention index of the corresponding link. The output is the link stability evaluation matrix and input into the frequency hopping decision engine.

[0031] In one embodiment of the present invention, UAV node 5 constructs a 512-byte dedicated probe frame carrying serial number #20250615 and dual timestamps (T1=10:00:00.500, T2=10:00:00.520) through the MAC layer and injects it into the evaluation link with node 7. After receiving the frame, the physical layer of node 7 counts 12 bit errors (instantaneous bit error rate 2.34×10⁻) based on FEC decoding. 5 ), record the reception completion time T3 = 10:00:00.560; the network layer matches the sequence number and calculates the end-to-end delay 60ms (T3-T1), the air interface delay 40ms (T3-T2), and generates a link quality tuple (bit error rate 2.34×10⁻ 5 , transmission delay 60ms, and air delay 40ms. Within a 30-second evaluation window, data from multiple links between node 5 and nodes 2, 7, and 9 is aggregated. The following calculations are performed on the 5-7 link: a bit error fluctuation factor of 0.15 (coefficient of variation), a delay jitter coefficient of 25ms (95th percentile range), and a channel contention index (number of times air delay exceeds 35ms) calculated. Finally, a stability evaluation matrix ([0.15, 25, 3]) is constructed with link ID "5-7" as a row. This matrix is ​​input into the frequency hopping decision engine to optimize the switching strategy for the main frequency of CH1.

[0032] In summary, the present invention realizes multi-dimensional real-time monitoring and evaluation of wireless link quality, can accurately obtain key indicators such as physical layer bit error rate, end-to-end transmission delay and air interface delay, and comprehensively calculate bit error fluctuation, delay jitter and channel contention, thereby providing comprehensive and accurate link stability evaluation data for the frequency hopping decision engine, effectively supporting intelligent frequency switching and communication quality optimization of Mesh networks in dynamic environments.

[0033] Preferably, the S110 is specifically: Extracting the historical bit error fluctuation factor and delay jitter coefficient of each link in the link stability evaluation matrix, calculating the time-attenuated weighted average of the historical bit error fluctuation factor and delay jitter coefficient, and generating a stability degradation gradient that reflects the long-term evolution trend of link quality; Aligning the interference intensity distribution feature in the current spectrum fingerprint feature with the stability degradation gradient in the time domain, establishing a mapping relationship between interference pulse energy and bit error fluctuation factor, and generating a frequency band vulnerability coefficient that characterizes the interference sensitivity of a specific frequency band; For the main frequency switching trigger threshold in the inter-frequency strategy table, the historical deterioration rate of the main frequency switching trigger threshold is determined based on the stability degradation gradient. Combined with the real-time error fluctuation factor of the current link, the deterioration margin factor is dynamically corrected through the proportional-integral regulator to output a dynamic switching threshold that is adaptive to changes in link quality; Based on the frequency band vulnerability coefficient, the frequency points that maintain the minimum protection interval with the frequency band where the interference pulse whose interference intensity value exceeds the preset interference threshold is located are screened out from the candidate frequency set, and the delay jitter coefficient corresponding to the link quality history data of the frequency point is extracted and the inverse is taken to generate the delay stability weight; at the same time, the frequency of occurrence of the frequency point in the heterodyne strategy table of the neighboring node is counted to generate the topology compatibility weight; The delay stability weight and the topology compatibility weight are linearly combined and normalized to obtain the comprehensive priority score of each alternative frequency. The fixed threshold in the original inter-frequency strategy table is replaced by a dynamic switching threshold, and the alternative frequency list is reordered according to the comprehensive priority score. The two highest-scoring backup frequencies are retained and their switching hysteresis intervals are updated to form an optimized inter-frequency strategy table.

[0034] Taking the drone cluster city inspection as an example, drone No. 5 extracts the historical stability data of the link with node No. 7 (the average error fluctuation factor in the past 10 minutes is 0.13, and the delay jitter coefficient is 22ms), and calculates the time-attenuation weighted average to generate the stability degradation gradient (deterioration rate +0.02 / minute). Combined with the 5.810GHz interference pulse (-65dBm) in the current spectrum fingerprint, a mapping relationship between the interference energy in this frequency band and the error fluctuation is established, and the vulnerability coefficient of the CH1 frequency band is obtained as 0.78. For the original inter-frequency strategy table, the main frequency CH1 switching threshold is 1.8×10⁻ 5 , based on the degradation gradient, the historical deterioration rate is determined to be 0.3×10⁻ 5 / minute, combined with the real-time error fluctuation factor of 0.18, the degradation margin factor is dynamically corrected from 1.8 to 2.1 through the PI regulator, generating a dynamic switching threshold of 2.1×10⁻ 5 From the candidate frequency set, we screened for CH5 / CH6 with a distance greater than 40 MHz from the interference band. We extracted the inverse of CH5's historical delay jitter coefficient to obtain a delay stability weight of 0.85. We also calculated the frequency of occurrence in the neighboring node strategy table (node ​​7 uses CH5) to obtain a topology compatibility weight of 0.75. After weighted normalization, CH6's comprehensive score of 0.82 surpassed CH5's 0.78. Finally, we updated the main frequency switching threshold to 2.1×10⁻. 5 , the order of alternative frequencies is adjusted to CH6 (level 1) and CH5 (level 2), and the hysteresis interval is synchronously updated to [1.8×10⁻ 5 ,2.3×10⁻ 5 ], forming an optimization strategy table.

[0035] In summary, the present invention realizes dynamic optimization and adjustment of the heterogeneous frequency strategy table by comprehensively analyzing the link stability degradation trend, interference sensitivity characteristics and network topology compatibility. It can adaptively correct the frequency switching threshold and alternative frequency priority, effectively improving the system's spectrum decision-making accuracy in time-varying interference environments. At the same time, by establishing a frequency band vulnerability assessment model and a weight fusion mechanism, it ensures that the selected frequency has both anti-interference capability and network coordination, thereby enhancing the communication stability and spectrum resource utilization of the Mesh network in complex electromagnetic environments.

[0036] In this embodiment, the method for intelligent dynamic inter-frequency networking of the image transmission module MESH self-organizing network further includes the following steps: Based on the frequency domain range covered by the conflicting frequency set and the bandwidth parameters of each interference pulse in it, the center frequency and bandwidth of the chaotic seed frequency band outside the legal frequency pool are determined, and the phase space coordinates that are conjugate symmetric with the interference pulse distribution are generated. Extracting a dynamic noise floor value calculated in the spectrum fingerprint feature, and activating a nonlinear modulator when a peak intensity of an interference pulse in the conflicting frequency set (the peak intensity is defined in S102 as a relative intensity ratio of the peak energy of the interference pulse to the dynamic noise floor) exceeds a preset multiple threshold of the dynamic noise floor; Perform complex domain convolution operation on the activated interference pulse time domain signal and the Lorentz attractor basis function corresponding to the chaotic seed frequency band to generate a chaotic modulation signal with spectrum shifted. The Lyapunov exponent of the chaotic modulation signal is obtained in real time. If the Lyapunov exponent is greater than zero and the phase space trajectory divergence rate exceeds the attractor convergence threshold, the bifurcation parameter is adjusted to make the signal energy fall within the preset dissipation boundary of the strange attractor. The Kolmogorov entropy change rate of the chaotic frequency band output signal is monitored. If the entropy change is lower than the preset annihilation efficiency threshold, the conjugate symmetry parameters of the phase space coordinates are iteratively updated until the interfering harmonic component disappears.

[0037] For example, in a drone swarm urban inspection scenario, drone node 5 selects a chaotic seed frequency band (center frequency 5.450 GHz, bandwidth 15 MHz) outside the legitimate frequency pool (5.725-5.850 GHz) based on the conflicting frequency set (5.810-5.830 GHz, including 15 MHz bandwidth / 30 dB intensity interference pulses). This frequency band generates phase space coordinates (real part 5.810, imaginary part 5.450) that are frequency-conjugate symmetric with the interference pulse. When the interference pulse intensity is detected to be 30 dB (exceeding the 25 dB threshold of the dynamic noise floor of -95 dBm), the nonlinear modulator is activated, performing complex convolution of the interference time domain signal with the Lorentz attractor basis function, generating a chaotic modulation signal with its spectrum shifted to 5.450 GHz. The Lyapunov exponent was calculated in real time and the divergence rate of the phase space trajectory was monitored (0.5>threshold 0.3). The bifurcation parameter was adjusted to 32 so that the signal energy fell into the dissipation boundary. The Kolmogorov entropy change rate was monitored (0.01bit / s<threshold 0.05bit / s). After the conjugate symmetric imaginary part was iteratively updated to 5.442GHz, the interfering harmonics disappeared.

[0038] It should be noted that in order to solve the problem in the existing technology that strong interference pulses are difficult to completely eliminate, resulting in waste of spectrum resources, this embodiment converts the interference signal energy into a controllable chaotic signal through chaotic modulation and phase space transformation and guides it to a preset dissipation frequency band, thereby realizing the active elimination of high-intensity interference and the recycling of spectrum resources, which can effectively suppress the impact of strong interference on the communication system and improve spectrum utilization.

[0039] In this embodiment, the method for intelligent dynamic inter-frequency networking of the image transmission module MESH self-organizing network further includes the following steps: According to the node distribution density of the real-time interference topology map, the topological potential energy field gradient of each relay node on the candidate frequency is calculated to generate a topological edge cutting risk value that reflects the sensitivity of network segmentation. Extract the QoS level identifier and data packet life cycle from the transport layer service flow label, and calculate the service criticality vector based on the service flow path hop count; Performing tensor fusion on the topology edge cutting risk value and the business criticality vector, and superimposing the neighborhood node alternative frequency weight similarity index to form a three-dimensional arbitration decision matrix; Under the constraints of a three-dimensional arbitration decision matrix, with the objective functions of minimizing the probability of network segmentation and maximizing the critical service pass rate, the Nash equilibrium solution set of the frequency selection strategy is solved using non-cooperative game theory. The Nash equilibrium solution set is input into the quantum annealing optimizer, which simulates the quantum tunneling effect to escape from the local optimal solution and output the final arbitration strategy that takes into account both topological connectivity and business assurance.

[0040] For example, drone node 5 calculates the topological potential field gradient of neighboring nodes (2, 7, and 9) on candidate frequency CH6 based on the real-time interference topology map. If node 7 is located at a network cut-edge location (density gradient 0.38), a topological cut-edge risk value of 0.62 is generated. The transport layer HD image transmission service flow (QoS label 0x0A, lifecycle 120ms, path hop count 3) is extracted and the service criticality vector [0.92, 0.85] is determined. The cut-edge risk value of 0.62 is fused with the criticality vector tensor and superimposed with the neighboring node's weighted similarity of 0.75 for CH6, forming a three-dimensional arbitration decision matrix ([0.62, [0.92, 0.85], 0.75]). With the constraints of minimizing the network partition probability (objective function f1) and maximizing the image transmission service pass rate (f2), a non-cooperative game was used to solve the Nash equilibrium solution set {CH5: 0.82, CH6: 0.88}. A quantum annealing optimizer (initial temperature 128, annealing rate 0.95) was input, and quantum tunneling was used to escape the local optimum of CH5. The final arbitration strategy was output: CH6 (comprehensive score 0.88) was selected as the service guarantee frequency.

[0041] In summary, to address the risks of network segmentation and critical service interruption caused by frequency selection in dynamic mesh networks (critical services include core communication services that require high priority protection, such as drone cluster control instructions, real-time high-definition image transmission data, and emergency alarm signals), this embodiment constructs a three-dimensional decision-making model by integrating topology potential analysis, service criticality assessment, and neighborhood collaboration mechanisms. This achieves coordinated protection of network topology connectivity and service reliability, effectively avoiding the risk of network segmentation during networking, improving the transmission success rate of critical services, and ensuring efficient utilization of spectrum resources through global optimization decisions.

[0042] like Figure 3 As shown, the second aspect of the present invention discloses a MESH self-organizing network intelligent dynamic frequency networking system for an image transmission module, which is applied to any of the above-mentioned methods for MESH self-organizing network intelligent dynamic frequency networking for an image transmission module, and the system includes: Distributed RF sampling unit 1, integrated with a broadband tunable RF front end, supports simultaneous real-time spectrum scanning within a preset frequency band, acquires raw time-domain signal data, and generates spectrum energy distribution maps based on fast Fourier transform; Spectrum fingerprint analysis module 2, connected to the distributed RF sampling unit, for calculating the dynamic noise floor, interference pulse characteristics and channel occupancy mode, and outputting structured spectrum fingerprint characteristics; Neighborhood interference topology construction module 3 receives the interference source feature table of neighboring nodes through a dedicated control channel, corrects the interference intensity value in combination with the signal attenuation factor, and generates a real-time interference topology map; A frequency hopping decision engine 4, connected to the conflicting frequency evaluation module, is used to screen a candidate frequency set from a preset legal frequency pool, calculate a frequency availability coefficient, and dynamically generate an inter-frequency strategy table including a primary and backup frequency and a switching threshold; Network topology management unit 5, supporting 2 to 264 nodes; 1 hop, 2 hops, 3 hops, 4 hops or unlimited hops; and supports software distance limit, up to 500KM; Multi-protocol compatible interface 6 supports hardware adaptation and software protocol stack expansion with the image transmission module.

[0043] The above are only specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered by the scope of protection of the present invention.

Claims

1. A method for intelligent dynamic frequency-interconnected networking of a MESH self-organizing network of an image transmission module, characterized in that: The following steps are involved: S102: Scanning the original time domain signal data of the preset frequency band of each target node in real time to generate spectrum fingerprint features including interference intensity distribution and channel occupancy pattern; S104: Based on the spectrum fingerprint features and in combination with the real-time interference topology map constructed through interaction between neighboring nodes, a conflicting frequency set to be avoided is identified through multi-dimensional weighted evaluation; S106: Dynamically generate an inter-frequency strategy table including primary and backup frequencies and switching thresholds using a frequency hopping decision engine based on the conflicting frequency set and a preset legal frequency pool, and synchronize the inter-frequency strategy table to associated relay nodes; S108: Collecting physical layer bit error rate and network layer delay data through a cross-layer frame detection mechanism, generating a link stability evaluation matrix and feeding it back to the frequency hopping decision engine; S110: Dynamically modifying the frequency switching threshold and the alternative frequency weight in the inter-frequency strategy table based on the historical sequence of the link stability evaluation matrix and the environmental change parameters.

2. The method for intelligent dynamic frequency-interconnected networking of a MESH self-organizing network of an image transmission module according to claim 1, characterized in that: The S102 is specifically as follows: Each target node uses a distributed RF sampling unit to synchronously sample the preset frequency band within a preset time window to obtain the original time domain signal data; Perform fast Fourier transform on the original time domain signal data to obtain the instantaneous signal energy value of each discrete frequency point and generate a spectrum energy distribution diagram; Based on the spectrum energy distribution diagram, the average background noise energy in the noise assessment sub-band is counted as the dynamic noise floor, and the dynamic interference determination threshold is set according to the noise floor; Scan the entire frequency band, identify areas where energy continuously exceeds the dynamic interference determination threshold as interference pulses, and record the center frequency, bandwidth, and peak energy of each interference pulse; For each interference pulse, the relative intensity ratio of its peak energy to the dynamic noise floor is calculated as the interference intensity value; Combine the center frequency, bandwidth and interference intensity values ​​of all interference pulses and map them to frequency domain coordinates to generate interference intensity distribution characteristics; Split the spectrum energy distribution diagram according to preset channel rules; The energy level of each channel is continuously monitored. If there is an interference pulse or the average energy exceeds the busy / idle threshold, the channel is marked as "busy", otherwise it is "idle". The start and end times of the state changes are recorded to form an occupancy event list. The occupancy rate, average occupancy time, and idle-busy transition frequency of each channel are calculated to generate a channel occupancy pattern. The interference intensity distribution characteristics are associated with the channel occupancy pattern and fused into a structured data object, which is output as the spectrum fingerprint feature.

3. The method for intelligent dynamic frequency-interconnected networking of a MESH self-organizing network of an image transmission module according to claim 1, characterized in that: The S104 is specifically as follows: Extract the interference intensity distribution characteristics from the spectrum fingerprint characteristics of the local node, analyze the center frequency, bandwidth and interference intensity value of each interference pulse, and generate a local interference source feature table. At the same time, receive the interference source feature table broadcast by the neighboring node through the dedicated control channel to form a set of neighboring interference sources. Determine a signal attenuation factor of a neighboring interference source on the current node based on the geographical location relationship between the nodes, and correct the interference intensity value in the neighboring interference source characteristic table based on the attenuation factor to generate an equivalent interference intensity value; The corrected neighboring interference sources and the local node interference sources are integrated according to the frequency domain coordinates and mapped into a real-time interference topology map with frequency as the horizontal axis and equivalent interference intensity as the vertical axis; Read the equivalent interference intensity value corresponding to each frequency point in the real-time interference topology map as the basic conflict weight; Correlate the channel occupancy patterns in the spectrum fingerprint features to extract the occupancy rate and idle-busy transition frequency of the channel to which the corresponding frequency belongs, thereby determining the channel instability coefficient; at the same time, count the number of times the corresponding frequency point appears in the neighborhood interference source feature table to generate the neighborhood conflict density value; The basic conflict weight, channel instability coefficient and neighborhood conflict density value of each frequency point are weighted to obtain the spectrum conflict risk value of each frequency point; the frequency points whose spectrum conflict risk value exceeds the preset risk threshold are screened, and the channel bandwidths to which they belong are merged to form a set of conflict frequencies that need to be avoided.

4. The method for intelligent dynamic frequency-interconnected networking of a MESH self-organizing network of an image transmission module according to claim 1, characterized in that: The S106 is specifically as follows: Eliminate all frequency points and associated channels covered by the conflicting frequency set from a preset legal frequency pool to generate a candidate frequency set; For each frequency point in the candidate frequency set, calculate its minimum frequency domain separation distance from the conflicting frequency set, and combine the average bit error rate performance of this frequency point in the historical link stability evaluation matrix to generate the frequency availability coefficient; Sort the frequency availability coefficients of each frequency point in the candidate frequency set in descending order, and select the frequency point with the highest ranking as the primary frequency; Among the remaining frequency points, based on the preset frequency band dispersion constraint rules, two frequency points that meet the constraints and have the highest frequency availability coefficients are selected as the primary backup frequency and the secondary backup frequency; The real-time physical layer bit error rate baseline value collected by the cross-layer frame detection mechanism is associated with the historical average bit error rate corresponding to the primary frequency and multiplied by the preset degradation margin factor to generate the primary frequency switching trigger threshold. Based on the predicted fluctuation range of the link quality of the primary backup frequency, the switching hysteresis interval between it and the primary frequency is set. The main frequency, two-level backup frequencies, main frequency switching trigger threshold and switching hysteresis interval are integrated into a structured inter-frequency strategy table; the inter-frequency strategy table is encrypted and encapsulated into a strategy update instruction frame through a dedicated control channel and broadcast to the associated relay nodes in real time.

5. The method for intelligent dynamic frequency-interconnected networking of a MESH self-organizing network of an image transmission module according to claim 1, characterized in that: The S108 is specifically as follows: A dedicated probe frame carrying a unique sequence number and dual timestamps is constructed at the media access control layer and injected into the transmit queue of the link to be evaluated. The dual timestamps include a transmit timestamp T1 and a scheduling timestamp T2. The scheduling timestamp T2 records the system time when the frame enters the physical layer buffer. When the receiving end's physical layer parses the arriving probe frame, it counts the original number of errored bits of the frame based on the forward error correction decoding result, and calculates the instantaneous bit error rate based on the probe frame length; it synchronously records the actual reception completion timestamp T3 of the frame, and extracts the sequence number to generate the original errored sample; The network layer captures the reception event of the probe frame and matches T1 and T2 registered by the sender according to the sequence number; calculates the difference between T3 and T1 to generate the end-to-end transmission delay, and calculates the difference between T3 and T2 to generate the air interface delay, forming a delay vector; Using the sequence number as a key, the original bit error samples corresponding to the same probe frame are bound to the delay vector to generate a link quality tuple including the bit error rate, transmission delay, and air delay; Within a preset evaluation window, link quality tuples of multiple links are aggregated. For each link, the coefficient of variation of the bit error rate is calculated as the bit error fluctuation factor, the percentile range of the transmission delay is calculated as the delay jitter coefficient, and the number of times the air delay exceeds the threshold is counted to generate the channel contention index. A two-dimensional matrix is ​​constructed with link IDs as rows and stability indicators as columns. Each row is filled with the bit error fluctuation factor, delay jitter coefficient, and channel contention index of the corresponding link. The output is the link stability evaluation matrix and input into the frequency hopping decision engine.

6. The method for intelligent dynamic frequency-interconnected networking of a MESH self-organizing network of an image transmission module according to claim 1, characterized in that: The S110 is specifically as follows: Extracting the historical bit error fluctuation factor and delay jitter coefficient of each link in the link stability evaluation matrix, calculating the time-attenuated weighted average of the historical bit error fluctuation factor and delay jitter coefficient, and generating a stability degradation gradient that reflects the long-term evolution trend of link quality; Aligning the interference intensity distribution feature in the current spectrum fingerprint feature with the stability degradation gradient in the time domain, establishing a mapping relationship between interference pulse energy and bit error fluctuation factor, and generating a frequency band vulnerability coefficient that characterizes the interference sensitivity of a specific frequency band; For the main frequency switching trigger threshold in the inter-frequency strategy table, the historical deterioration rate of the main frequency switching trigger threshold is determined based on the stability degradation gradient. Combined with the real-time error fluctuation factor of the current link, the deterioration margin factor is dynamically corrected through the proportional-integral regulator to output a dynamic switching threshold that is adaptive to changes in link quality; Based on the frequency band vulnerability coefficient, the frequency points that maintain the minimum protection interval with the frequency band where the interference pulse whose interference intensity value exceeds the preset interference threshold is located are screened out from the candidate frequency set, and the delay jitter coefficient corresponding to the link quality history data of the frequency point is extracted and the inverse is taken to generate the delay stability weight; at the same time, the frequency of occurrence of the frequency point in the heterodyne strategy table of the neighboring node is counted to generate the topology compatibility weight; The delay stability weight and the topology compatibility weight are linearly combined and normalized to obtain the comprehensive priority score of each alternative frequency. The fixed threshold in the original inter-frequency strategy table is replaced by a dynamic switching threshold, and the alternative frequency list is reordered according to the comprehensive priority score. The two highest-scoring backup frequencies are retained and their switching hysteresis intervals are updated to form an optimized inter-frequency strategy table.

7. The method for intelligent dynamic frequency-interconnected networking of a MESH self-organizing network of an image transmission module according to claim 2, characterized in that: The channel busy / idle threshold is set based on a dynamic noise floor.

8. The method for intelligent dynamic frequency-interconnected networking of a MESH self-organizing network of an image transmission module according to claim 4, characterized in that: The preset frequency band dispersion constraint rule requires that the alternative frequencies and the primary frequency and each other maintain a minimum frequency band interval.

9. An intelligent dynamic inter-frequency networking system for a MESH self-organizing network of an image transmission module, applied to an intelligent dynamic inter-frequency networking method for a MESH self-organizing network of an image transmission module according to any one of claims 1 to 8, characterized in that: The system comprises: Distributed RF sampling unit, integrated with a wideband tunable RF front end, supports simultaneous real-time spectrum scanning within a preset frequency band, acquires raw time-domain signal data, and generates spectrum energy distribution maps based on fast Fourier transform; A spectrum fingerprint analysis module, connected to the distributed RF sampling unit, is used to calculate the dynamic noise floor, interference pulse characteristics and channel occupancy mode, and output structured spectrum fingerprint characteristics; The neighborhood interference topology construction module receives the interference source feature table of the neighboring nodes through a dedicated control channel, corrects the interference intensity value based on the signal attenuation factor, and generates a real-time interference topology map; A frequency hopping decision engine, connected to the conflicting frequency evaluation module, is used to screen a candidate frequency set from a preset legal frequency pool, calculate a frequency availability coefficient, and dynamically generate an inter-frequency strategy table including a primary and backup frequency and a switching threshold; Network topology management unit, supporting 2 to 264 nodes; 1 hop, 2 hops, 3 hops, 4 hops or unlimited hops; and supports software distance limit, up to 500KM; Multi-protocol compatible interface supports hardware adaptation and software protocol stack expansion with image transmission modules.

Images