171results about How to "Avoid detection" patented technology

The invention discloses a self-adaptive resource scheduling method for improving cluster reconstruction efficiency, and relates to the field of resource scheduling distribution, and the method comprises the steps: monitoring the resource use data of each node in a cluster in real time through a monitoring agent, and obtaining the health state of the node through a heartbeat mechanism; collecting input task requests, predicting the resource demand quantity of each task, and performing preliminary queue processing on the tasks to obtain a task queue group; setting a resource competition constraint according to the task queue group and the resource use data and the health state of each node, obtaining dynamic resource scheduling corresponding to each task by using an optimization algorithm, and allocating the tasks to each node; a resource scheduling dynamic graph is constructed, a visual interface is set, and the resource scheduling dynamic graph is arranged on the visual interface and used for displaying resource scheduling, providing a user-friendly interface and realizing high-efficiency scheduling of adaptive resources in a cluster reconstruction process.

The invention relates to the field of elevator door change gear detection, in particular to an elevator door change gear detection device, which solves the problem that a door change gear can be automatically replaced while the elevator door change gear detection device is suitable for simultaneous detection of various door change gears with different sizes, and comprises a cabinet body and a driving motor, and an output shaft is fixedly installed at the output end of the driving motor, a plurality of driving wheels are fixedly installed on the outer wall of the output shaft, and the driving wheels are installed at equal intervals. According to the device, the function of automatically adjusting the height of the change gear seat can be achieved, the device can be conveniently suitable for door change gears with different sizes and diameters, the device can conveniently detect the door change gears with different sizes at the same time, a replacement box of the device can achieve automatic replacement when the door change gears descend after detection is completed, and the detection efficiency is improved. Manual operation is not needed, and the detection efficiency of the device can be improved on the premise of ensuring safety.

The invention discloses a supplier quality evaluation method, and relates to the technical field of data quality analysis, and the method comprises the steps: obtaining a data initial quality evaluation label of each supplier; counting an uploaded information quality processing comprehensive data quality evaluation value of each type of suppliers; determining an initial quality detection constraint period of each type of suppliers and performing quality tracking detection; obtaining quality tracking detection data, and analyzing and processing to obtain a quality tracking monitoring result; according to the method, the data quality of the data uploaded by the suppliers is analyzed, the availability of the quality evaluation data is ensured, the suppliers whose initial quality evaluation is unqualified are screened, and the quality of the suppliers is evaluated to be qualified. Quality evaluation resources are concentrated on suppliers whose initial quality evaluation is qualified, customized quality tracking detection is carried out on each type of suppliers, the quality tracking detection efficiency and result are improved, and the quality detection constraint period is dynamically adjusted in time.

The invention discloses a power transmission line foreign matter recognition method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining a to-be-recognized first power transmission line picture and a first cue word used for representing information content needing to be recognized from the first power transmission line picture; wherein the information content comprises the type and position information of the foreign matter; inputting the first power transmission line picture and the corresponding first cue word into a multi-modal model, so that the multi-modal model generates the type and position information of the foreign matter in the first power transmission line picture; the accuracy of foreign matter identification of the power transmission line can be improved.

The invention relates to the technical field of rubber tire detection, in particular to an air tightness detection device for rubber tire production and a detection method thereof. Comprising a detection platform, and a telescopic groove is formed in the top of the detection platform; and a tire supporting mechanism movably penetrates through the telescopic groove. The tire limiting mechanism is controlled to limit and fix a rubber tire, then the tire supporting mechanism is controlled to be separated from the rubber tire, the detection marking mechanism is controlled to wrap a part of the rubber tire, and then the tire limiting mechanism is controlled to drive the rubber tire to rotate; through the arrangement of the detection marking mechanism, the detection marking mechanism can mark the air leakage position while detecting the surface in the rotation process of the rubber tire, the problems that manual detection wastes time and omission is likely to happen are solved, the detection work efficiency is improved, the air leakage point can be accurately marked, and the detection efficiency is improved. The problem that follow-up repair is inconvenient due to the fact that no mark exists after existing liquid detection is dried is solved, and the using effect of the detection device is improved.

The utility model discloses a pontoon type acoustic Doppler flow velocity profiler, which relates to the field of acoustic Doppler flow velocity profilers, and comprises a pontoon main body and an acoustic Doppler flow velocity profiler main body, and two electric push rods are arranged on the acoustic Doppler flow velocity profiler main body. Movable rings are installed at the output ends of the two electric push rods through piston rods, a plurality of connecting blocks are arranged on the outer wall of the acoustic Doppler flow velocity profiler body, stop levers are connected to the connecting blocks through rotating shafts, and push plates are connected between the stop levers and the movable rings through connecting shafts. According to the acoustic Doppler flow velocity profiler, the blocking rod and the protruding block are arranged, the blocking rod can protect the acoustic Doppler flow velocity profiler body and prevent objects from impacting the acoustic Doppler flow velocity profiler body, the protruding block is arranged so that the plastic bag can be conveniently hung, and the plastic bag in water is prevented from being wrapped on the outer wall of the acoustic Doppler flow velocity profiler body; therefore, the acoustic Doppler flow velocity profiler main body is prevented from detecting the water body.

The embodiment of the invention discloses a power converter and a control method thereof, relates to the technical field of energy, and solves the problem of how to accurately judge whether an arc fault breaker works normally or not. Specifically, in the power converter, a controller is used for outputting a pulse signal. The noise signal conditioning circuit is used for receiving the pulse signal and outputting a noise signal according to the pulse signal. The current transformer is used for receiving and sampling the noise signal so as to output a sampling signal. The sampling signal conditioning circuit is used for receiving the sampling signal and outputting a conditioning signal according to the sampling signal. The controller is further used for controlling the power conversion circuit to start working under the condition that the absolute value of the difference between the frequency of the conditioning signal and the frequency of the pulse signal is smaller than or equal to a frequency threshold value and the gain of the conditioning signal is larger than or equal to a gain threshold value corresponding to the frequency of the pulse signal.

The invention relates to the technical field of pressure leakage stoppage of pressure-bearing equipment, and particularly provides a pressure leakage stoppage clamp for a pressure-bearing equipment flange and a preparation method of the pressure leakage stoppage clamp, and the pressure leakage stoppage clamp comprises a first clamp and a second clamp which are detachably connected to a flange plate gap and a cylindrical surface and are of an integrally-formed elastic deformation structure; each of the first fixture and the second fixture comprises a fixture body, a fixture boss and a fixture bevel edge, and when the first fixture and the second fixture are mounted and fastened on the flange, the fixture bevel edges are stressed to elastically expand and deform, form circumferential interference contact with the cylindrical surface and generate axial component force to act on the side surface of the flange; the boss is tightly attached to the surface of the flange plate gap in the direction of the flange plate gap through the generated counter-acting force, and the boss of the clamp and the side surface of the flange plate gap are in installation interference fit, so that the flange plate gap is sealed, and a sealed cavity is formed; and then glue is injected to realize pressure sealing of the flange of the pressure-bearing equipment.

The invention relates to the technical field of industrial automatic coating, in particular to a groove frame paint spraying system which comprises a rotary carrying table device, a paint spraying device and a paint spraying device. The visual system is used for photographing the to-be-painted groove frame and determining a spraying area and a cover area on the to-be-painted groove frame according to image information, obtained through photographing, of the to-be-painted groove frame; and the spraying robot is used for spraying an isolation layer material to the cover area and spraying anti-corrosion paint to the spraying area after a preset first time period of spraying of the isolation layer material is completed in the cover area. The spraying area and the cover area of the groove frame are recognized through the visual system, accurate area segmentation is achieved, the spraying robot can accurately control the spraying positions of an isolation layer and anti-corrosion paint, and the phenomena of paint waste and mistaken spraying are avoided. The anti-corrosion paint is sprayed after the isolation layer is sprayed, it is guaranteed that the specific area of the surface of the groove frame is not affected by the anti-corrosion paint, meanwhile, other areas are protected, and the corrosion resistance and the surface protection effect of the groove frame are enhanced.

The invention discloses a backdoor attack method for personalized federal learning based on feature aggregation, and belongs to the technical field of backdoor attack. The size of the designed back door trigger is equal to that of an input image, and each pixel value of the trigger is constrained. According to the design, detection based on image tampering defense can be effectively avoided, so that the backdoor trigger is more difficult to find visually. Secondly, training a backdoor trigger and a backdoor model by adopting an intra-class attack strategy; in the process, a backdoor trigger is trained by using a feature distribution alignment method, so that backdoor data generated by a malicious client is as close as possible to data distribution of other normal clients, and a federated defense mechanism based on model parameter clustering detection is avoided. And finally, all malicious clients cooperatively train a backdoor trigger. Feature aggregation of the backdoor trigger is carried out between malicious clients, data advantages mastered by an attacker are fully utilized, the generalization ability of the backdoor trigger is improved, and the backdoor trigger can still keep effective under wide input conditions.

The utility model discloses a wafer non-contact positioning platform device, which comprises a bottom plate, a first vertical plate, a second vertical plate, a first transverse plate and a second transverse plate, a centering platform is rotatably connected above the first transverse plate, a motor is connected below the middle of the centering platform, a lifting mechanism is arranged on the bottom plate, and a laser displacement sensor is arranged on the second vertical plate; the centering platform comprises a base plate, a first connecting plate, a first positioning block, a second connecting plate and a second positioning block, and the ends, away from the base plate, of the first positioning block and the second positioning block are each of a step-shaped structure. The beneficial effects of the utility model are that edge searching or V-Notch port detection of the wafer can be realized, the orientation of the wafer can be rapidly determined, and the position of the wafer can be rapidly positioned; the positioning platform device is only in contact with the edge of the wafer in the orientation process and is not in direct contact with a wafer chip, so that the scratch and abrasion risks of the wafer chip can be effectively reduced, the wafer is not physically damaged, and the orientation quality of the wafer is ensured.

The invention discloses a transmission scheme design and performance optimization method of a symbiotic radio system based on an intelligent reflection surface. In the scheme, the intelligent reflection surface assists the multi-antenna base station to communicate with a main user, and meanwhile, the intelligent reflection surface serves as a backscattering device to generate a secondary signal and transmit information to a secondary user. The intelligent reflection surface controls the on-off of the reflection unit in an on-off keying mode, modulation and scattering of base station reflection signals are achieved, and a secondary user carries out secondary symbol detection through an energy detection method. On the premise of ensuring the transmission rate of main user transmission, joint optimization is performed on a base station beam forming vector and a phase shift matrix of an intelligent reflection surface so as to minimize the error probability of scattering transmission. An original optimization problem is a non-convex problem and is converted and solved by adopting a positive semidefinite relaxation method, an alternating iterative optimization method and other methods. Compared with a conventional symbiotic radio transmission scheme, the scheme provided by the invention has a higher primary transmission rate and a lower secondary transmission bit error rate.

The invention provides a new energy automobile storage battery heat dissipation and fire extinguishing mechanism, and relates to the field of new energy automobiles. The new energy automobile storage battery heat dissipation and fire extinguishing mechanism comprises a first mounting box, a second mounting box, a U-shaped frame and a storage battery body, the upper surface of the first mounting box is fixedly connected with the second mounting box, and mounting blocks are fixedly connected to the front ends and the rear ends of the lower portions of the outer walls of the two sides of the first mounting box correspondingly; and two mounting holes are formed in the upper surface of each mounting block, a mounting plate is fixedly connected to the middle of the outer wall of the rear end of the first mounting box through a plurality of bolts, and mounting cavities are formed in the two sides of the middle of the outer wall of the rear end of the mounting plate. According to the heat dissipation and fire extinguishing mechanism for the storage battery of the new energy automobile, air cooling and heat dissipation of the storage battery body are achieved through the two servo motors provided with the heat dissipation fans, heat dissipated by the storage battery body is absorbed through the U-shaped frame, the heat conduction pieces and the heat conduction rods, and then the heat dissipation effect of the storage battery body is improved.

The invention belongs to the technical field of intelligent equipment fault detection, and discloses a multi-robot-based large-scale equipment abnormal component cooperative detection method. According to the large complex intelligent equipment abnormal component detection method based on the multi-robot and cloud edge collaborative architecture, clustering, an A * algorithm, shortest route distribution and a Kalman filtering fusion algorithm are utilized, and multi-robot rapid inspection and accurate abnormal component detection are achieved. Through multi-robot uniform distributed inspection, global three-dimensional coordinate system establishment, abnormal point secondary inspection, abnormal point clustering analysis and screening, task list distribution, observable path division and fitting and optimal detection scheme construction, multi-angle and multi-mode abnormal data are finally fused, and an accurate detection result is obtained. The multiple robots are evenly distributed around the large complex intelligent equipment at equal intervals for inspection, the detection efficiency is improved, the coverage range is wider, and abnormal information can be fed back in time.

The invention relates to the technical field of data processing, in particular to a blowing flow adjusting system in the metal powder machining process. The system comprises the steps of collecting flow data, obtaining a mutation factor of each data in a main body flow data sequence, and obtaining a noise factor of each data in the main body flow data sequence according to the mutation factor of each data in the main body flow data sequence; based on the noise factor, adaptively obtaining a judgment threshold value of a Mahalanobis distance of each datum in the main body flow data sequence; according to the judgment threshold value of the Mahalanobis distance of each data in the main body flow data sequence, noise data is obtained for smoothing, the denoising effect is improved, and the influence of abnormal flow data on noise data recognition is avoided.

This invention provides a valve internal leakage diagnostic device, including a valve body, a hollow support base integrally connected to the middle of the valve body, an internal sealing assembly inside the hollow support base, a rotating handwheel connected to the top of the internal sealing assembly, pipes connected to both sides of the valve body, flanges connected to the ends of the pipes, four sets of mounting holes evenly distributed on the flanges, an adjustable support assembly fitted on the outer wall of the pipes, and an internal leakage detection assembly connected to the side of the adjustable support assembly facing the outer wall of the pipes. This invention has the following advantages: its simple structure eliminates the need to disassemble the valve for inspection, and it can be directly installed outside the valve body, enabling the inspection of a large number of valves, thus avoiding the problem of some valves having internal leaks that are difficult to detect and thus causing leakage.

This application discloses a method, apparatus, air conditioning system, and storage medium for detecting the lid status of an electrical control box, relating to the field of air conditioning equipment technology. The method for detecting the lid status of the electrical control box includes: acquiring the opening degree of the throttling device, the rotational speed of the cooling fan, the frequency of the compressor, and the ambient temperature of the electrical control box when the electrical control box is in an overheated operating state; and determining the lid status of the electrical control box based on the opening degree of the throttling device, the rotational speed of the cooling fan, the frequency of the compressor, and the ambient temperature of the electrical control box. This method automatically detects the lid status of the electrical control box using the opening degree of the throttling device, the rotational speed of the cooling fan, the frequency of the compressor, and the ambient temperature of the electrical control box, thereby avoiding manual inspection of the lid and improving the efficiency of lid status detection.

The application relates to the technical field of surface defect intelligent detection, and discloses a surface defect detection method and device based on multi-view synthetic imaging. The method comprises the following steps: acquiring a multi-view synthetic image without dead angles of a detected target after quality evaluation; performing highlight suppression and shadow enhancement on the multi-view synthetic image to obtain an enhanced multi-view synthetic image; separating the enhanced multi-view synthetic image into corresponding multiple single-view images, respectively performing defect detection on each single-view image to obtain corresponding single-view defect detection results, and generating a final overall detection result based on the single-view defect detection results. Through multi-view synthetic imaging, image quality control, high-reflective surface enhancement processing and multi-view joint defect detection, the application can effectively solve the problems of light interference and defect missing detection in high-reflective surface detection, and can realize high-precision surface defect detection with no dead angle coverage in a single detection.

This application relates to the technical field of electrical testing and discloses an online four-wire integrated testing device. The testing device includes a frame, a moving mechanism, and a testing mechanism. The frame has a testing platform, the moving mechanism is mounted on the testing platform, and the product under test is placed on the horizontal moving end of the moving mechanism. The testing mechanism is located above the moving mechanism and includes a lifting assembly, several probe groups, and several electrical drive rods. The lifting assembly is mounted on the frame, and each electrical drive rod is detachably connected to the vertical drive end of the lifting assembly. The probe groups are slidably connected to the electrical drive rods to correspond to the test ports of the product under test. This application effectively improves the quality inspection efficiency of circuit board products during the manufacturing process, enabling rapid detection of multiple defects in the online production stage, avoiding the tediousness and inaccuracy of manual inspection, and reducing downtime in the production process.

The utility model discloses a sealing surface precision detection oil pipe joint detection device, especially in oil pipe joint detection technical field, including support plate, be provided with the flow guide detection subassembly on the support plate, the flow guide detection subassembly includes the first flow guide box of setting in support plate one side, one side of first flow guide box is provided with the second flow guide box. The utility model discloses through pump machine and pressure sensor cooperation, can accurate simulation downhole high pressure environment, and pressure regulation range adapts different specifications oil pipe joint detection demand, solves the problem that the existing device can not restore actual working condition, leads to the detection result and actual use scene disjunction, avoids the problem of the leakage of the detection, the misjudgment caused by the inaccurate simulation of the working condition.

Embodiments are directed to preventing sandbox environments from being detected by potentially malicious applications. To this end, execution of an application is monitored and information about the execution is provided to a reinforcement learning machine learning model. The model generates suggested modifications to be made to the executing application. The model is provided information indicating whether the application successfully executed and this information is used to train the model for other modifications. By modifying the execution of potentially malicious software during execution, detection of the sandbox environment can be prevented and better analysis of potentially malicious application features can be understood.

The utility model relates to the technical field of camouflage nets, and discloses a sandy soil camouflage toggle clip camouflage net structure which comprises a bottom net, a color strip layer is arranged at the top of the bottom net, a plurality of groups of camouflage strips are arranged on the color strip layer, a plurality of groups of toggle clips are arranged at the top of the bottom net, and the toggle clips are used for limiting and fixing the camouflage strips. According to the sandy soil camouflage toggle clip camouflage net structure, the camouflage strips on the colored strip layer can be fixed through the toggle clip piece, stable supporting and fixing points are provided for the camouflage strips, the toggle clip can tightly grasp the camouflage strips under various complex environment conditions no matter whether the camouflage strips are lightly fished by breeze or attacked by strong wind, the camouflage strips are effectively prevented from flying around, and the camouflage net structure is simple in structure and convenient to use. In this way, the camouflage net always keeps complete and stable appearance in the long-time use process, and the reliability and durability of the camouflage effect are greatly improved.

The wearable lumbar and dorsal muscle training detection device comprises side bandages arranged on the two sides of the waist, the side bandages are connected through elastic bands, and hard plates are fixed to the front ends of the side bandages. Extension plates are upwards arranged on the rear sides of the side bandages, and detection motor pieces corresponding to the erector spinalis are evenly installed on the inner sides of the extension plates and used for detecting the stress states of the left erector spinalis and the right erector spinalis in the training process. A tightness observation mechanism is arranged on the outer side of the hard plate on one side and comprises a fixed box, a sliding plate, a boss, a sliding groove, a spring and an indication line, and the wearing tightness degree is visually displayed through the position change of the sliding plate. A tightening mechanism is arranged on the outer side of the hard plate on the other side and comprises a supporting plate, a rotating shaft, a rotary knob, an inclined plate, a torsional spring and a wedge-shaped groove matched with the connecting belt, tightening and fixing of the connecting belt are achieved through a one-way limiting structure, and locking is controlled to be released through the rotary knob so as to achieve quick detaching.

This utility model relates to a mechanical comparator suitable for bearing testing, comprising a housing, a first positioning block fixedly mounted on the top of the housing by a first bolt and a second bolt, and a base plate fixedly connected to the bottom of the housing; an adjusting block fixedly mounted inside the housing, the adjusting block being connected to a first guide post, the first guide post being fitted with a first guide sleeve; the adjusting block being connected to a second guide post, the second guide post being fitted with a second guide sleeve; a fixing block being mounted on the upper end of the first and second guide posts; and a second positioning block being connected to the fixing block; the inner diameters of both the first and second positioning blocks are matched to the bearing assembly to be tested; an adjustment device is mounted on the base plate for adjusting the gap between the fixing block and the adjusting block. Thus, the inner holes of the first and second positioning blocks precisely match the outer ring of the bearing to be tested, and with the coaxiality control of the mandrel and measuring ring, the bearing is ensured to be subjected to uniform force, avoiding subjective errors caused by reliance on worker experience in traditional dial indicator testing.

The present invention discloses a bolt preload evaluation method and a dynamic parameter measurement device thereof, the method comprising: considering constant and dynamic parameters, establishing a bolt preload evaluation model based on the energy method; obtaining the geometric dimensions of the bolted structure, the warping displacement of the measuring point and the parameter data of the physical properties of the material; determining the angular displacement of the thin plate according to the warping displacement of the measuring point and its distance from the axis of the tightening bolt; determining the bolt preload based on the parameter data, the angular displacement and the evaluation model. In addition, the device comprises a measuring point warping displacement and a bolt preload measurement system. The technical solution provided by the present invention can improve the level of structural reliability design by effectively evaluating the bolt preload, provide a theoretical basis for the high-performance assembly of mechanical equipment and the improvement of service performance, and reduce maintenance costs.

The invention relates to the technical field of artificial intelligence, and provides an agent interaction-based document conversion method, which comprises the following steps of: splitting a target document according to pages, and distributing an agent for each split page; detecting the embedded resources of the distributed page by the intelligent agent which accepts the page distribution; if the agent accepting page allocation detects that the risk element exists in the allocated page, the risk element is removed, and the page with the risk element is converted into a page in a security format; if the agent receiving page allocation detects that an error exists in the allocated page, other agents are informed of the error, and the agents obtaining the error generate an error correction scheme of the error page through the shared knowledge base; error correction is carried out on the page with errors by adopting an error correction scheme, and the page after error correction is converted into a page in a safe format; and recombining the pages in the secure format into a complete document to be output. According to the technical scheme, the document conversion can be ensured to be safe, efficient and consistent.

The invention relates to the technical field of scanning devices, in particular to a quick response X-ray scanning device based on a micro-electro-mechanical chip, which comprises a detection table, a photoelectric sensor is arranged in the detection table, a plurality of positioning blocks are arranged on the detection table, and a table top of the detection table is used for bearing an object to be imaged; the X-ray source is arranged above the detection table and can emit X-rays to the direction of the detection table, and the X-ray source is connected with a movable support capable of driving the X-ray source to move; the detector is connected with the detection table and can receive the X-rays penetrating through the object to be imaged and convert the X-rays into digital signals; the detection table fixes the position of an object to be imaged through the positioning block, the image acquisition module acquires an image of the detection table, and the control module adjusts the position of the X-ray source according to the position of the object to be imaged on the image so as to completely scan the object to be imaged. By accurately positioning the position of the object to be imaged and standardizing the image of the object to be imaged, the accuracy and reliability of the whole scanning process are improved, and the scanning efficiency and the imaging quality are improved.

This invention relates to an active learning-based data-free black-box attack method and system based on multidimensional value assessment, belonging to the field of artificial intelligence security technology. This method constructs a pre-emptive "sample screening funnel," utilizing a local substitution model to perform multidimensional assessments of sample boundary approximation, information uncertainty, and geometric diversity before sending images to a commercial cloud API. Only high-value samples are selected for querying, thereby achieving low-cost, high-efficiency model theft and adversarial attacks. This invention ensures the diversity and training stability of data-free generated samples, significantly improves the transfer success rate of adversarial examples, and achieves "low-cost, low-risk" economical attacks. It has strong versatility and can be seamlessly integrated into various existing data-free attack frameworks, facilitating deployment and implementation in practical security assessment systems.

The utility model relates to the technical field of tea packaging equipment, in particular to a falling cover press-fitting mechanism. According to the technical scheme, the device comprises an automatic conveying belt used for conveying can covers; the supporting table is located on one side of the automatic conveying belt, multiple sets of supporting frames are arranged on the two sides of the supporting table, and a conveying belt used for conveying tank bodies is fixedly arranged between the supporting frames; the cover falling assembly is located between the conveying belt and the automatic conveying belt and transfers the can cover to the top of the can body; the cover pressing assembly is installed on the top of the supporting table and used for pressing a tank cover on the top of the tank body; and the detection assembly is located at the tail end of the supporting table and used for detecting the pressing state of the tank cover and whether the top of the tank body leaks the cover or not. The tea sealing cover can be placed accurately, a can body is prevented from being damaged when the cover is pressed, and the tea packaging efficiency and quality are improved.

The invention discloses a region detection model training method, a region detection method and a region detection device. The method comprises the following steps: acquiring a plurality of training samples, inputting a target image in the training samples into a preset region detection model, detecting a drivable region in the target image through the region detection model, outputting to obtain a prediction image, extracting the edge of the predicted drivable region in the prediction image to obtain a first edge image, and extracting the edge of the travelable region in the annotated image to obtain a second edge image, determining an edge loss function value according to the first edge image and the second edge image, and adjusting the model parameters of the region detection model under the condition that the edge loss function value does not meet the training stop condition to obtain a travelable region. And training the region detection model after parameter adjustment by using the training sample until a training stop condition is met, thereby obtaining a trained region detection model. In this way, the area detection model can detect the drivable area with the more accurate edge.