186results about How to "Improve repair accuracy" patented technology

The invention relates to a data cleaning and recovering algorithm of a floating car related to a multi-threshold space. The algorithm comprises the processes of: merging data of a road section into a data matrix, wherein the data is collected by the floating car within a plurality of days; further removing abnormal data in the data matrix according to a three-sigma filter rule after taking a day as a unit to primarily screen missing data and abnormal data in the data matrix; then recovering the missing data by respectively combining space similar characteristics of the road section by a weighted average method and an exponential smoothing method according to the type of the missing data; and finally, recovering high-frequency noise disturbance in data of the floating car by a principal component rebuilding method, thereby fully excavating regularity speed information of the floating car under the condition of not increasing the quantity of the floating car and additional handling equipment, and further, effectively improving the data quality of the floating car. The data cleaning and recovering algorithm can be used for cleaning and recovering work of regularity data of the floating car, and can provide reliable data for subsequent transportation planning and traffic identification.

The invention relates to a device integrating ultrasound assisted electrospark deposition repair and ultrasonic polishing and a method thereof. The device comprises a gas feeding device, a power source and a control box (34), a connection (35), an electrode assembly (37) and a workpiece clamp head (39). By using the combination of the electrospark deposition technique and the ultrasonic mechanical vibration rubbing effect, the repair of the surface defect of a mould can be realized, and the surface of the mould can be mechanically polished by using the ultrasonic vibration energy after the repair is completed. In the invention, the electrospark deposition technique and the ultrasonic polishing technology are comprehensively utilized, so that the repair cost is low while repair precision is high. The combination processing can not only repair the invalid surface material of the mould with high efficiency, but also suit the occasions when major repair equipment is not available. The method for integrating ultrasound assisted electrospark deposition repair and ultrasonic polishing has the advantages of simple operation, convenience and practicality.

The invention provides an automation straightener for a drilling tool, which comprises a base, a pressure machine, a roller, a lifting device, a movable cushion block, a detecting trolley, and a control module, wherein the base is used for supporting; the pressure machine is used for pressing against the straightening drilling tool and is movably arranged on the upper part of the base; the roller is used for rotating the drilling tool, and one end of the roller is arranged on the base in a stretchable connection way; the lifting device is used for jacking the drilling tool, and one end of the lifting device is fixedly arranged on the base and positioned at the outer side of the roller; the movable cushion block is used for supporting the drilling tool, and one end of the movable cushion block is movably arranged on the base; the detecting trolley is used for detecting shape and position data of the drilling tool and is movably arranged on the base, and transmits the data to the control module; and the control module receives the data transmitted by the detecting trolley, and controls the actions of the pressure machine, the roller, the lifting device, the movable cushion block and the detecting trolley according to the data and a preset program. The automation straightener for the drilling tool realizes automatic detection and automatic control when the drilling tool is repaired by using the detecting trolley and the control module, thereby largely reducing human intervention, realizing the automation for drilling tool repair and achieving high accuracy of drilling tool repair, rapid speed, and high stability.

A method for determining a root cause of a problem in a multiple-element system includes correlating an incoming alarm signal with an element x in a multiple-element system and accessing an implication list comprising a list of all elements upstream of element x. At least one element on the implication list is weighted with data relating to the at least one element. Taking into account the weighting step, a probability is calculated that an element on the implication list comprises a most-probable root cause of the subsequent alarm signal. The identified most-probable root cause of the incoming alarm signal is output.

The invention discloses a mould three-dimensional rapid forming and repairing method based on reverse engineering. Three-dimensional reverse engineering is carried out on a faulted mould through the reverse engineering technology, comparison calculation of a faulted mould model and a standard mould model is finished in combination with a CAD technology, and recognition and obtaining of a CAD model in a to-be-repaired region of the mould are realized; slicing processing is carried out on the CAD model in the to-be-repaired region of the mould through the layering thought of the rapid forming technology, then three-dimensional rapid forming and repairing are carried out on the to-be-repaired region of the mould step by step and layer by layer through the argon arc welding technology, finally proper after-treatment and machining are carried out on the repaired region, and rapid repairing and re-manufacturing of the mould are finished. Accurate control over the repairing layer thickness and repairing amount of the mould is realized, the repairing after-treatment links are effectively reduced and even eliminated, the repairing accuracy and quality of the mould are improved, and an effective means with low cost, high accuracy and high efficiency is provided for on-site repairing and re-manufacturing of the mould.

The present invention relates to the quantitative remote sensing data processing technology, and discloses space-time integrated fusion method of remote sensing land surface temperature data. The method comprises first pre-processing the remote sensing land surface temperature data, then solving land surface temperature diurnal change model parameters by using multidate temperature data and corresponding time, thus to realize space seamlessness, screening similar picture elements window by window on polar orbit satellite land surface temperature image data, and in combination with weight factors such as time difference, scale difference, similarity difference and Euclidean distance composed by an multidate motionless satellite and the above model, commonly establishing a space-time integrated fusion model and solving a high temporal-spatial resolution land surface temperature dataset for generating space seamlessness. According to the method provided by the present invention, space-time complementary information of land surface temperature data of the polar orbit satellite and the multidate motionless satellite and diurnal change rules of motionless satellite land surface temperature are fully utilized, and no auxiliary data is required, such that the method is easy to realize, high in practical value and applicable to remote sensing land surface temperature data operational running with seamless space and time.

The invention discloses a method for beautifying composition of an image based on main body adjustment. The method for beautifying the composition of the image based on main body adjustment comprises the following steps that firstly, main body area detection is conducted on the image, so that a main body detection image is obtained; secondly, morphologic processing and binarization are conducted on the main body detection image, so that a main body mask binary image is obtained; thirdly, a main body is removed from the image through the main body mask binary image, zooming is conducted on the main body according to the optimal main body ratio rule, and the size of the main body is adjusted; fourthly, a blank area, generated after the main body is removed, of the image is filled in through the image restoration algorithm based on a sample, so that a background image is obtained; fifthly, according to the rule of thirds in the image composition aesthetics, the zoomed main body is moved to the background image, so that the beautified image is obtained. The invention further discloses a system for beautifying the composition of the image based on main body adjustment. The method and system for beautifying the composition of the image based on main body adjustment achieve automatic beautification of the composition of the image and have the advantages of being high in operating speed, strong in feasibility and good in automatic beautification effect.

The invention relates to a portable ultrasonic auxiliary spark sedimentation repairing and polishing integrated device and a process thereof. The device comprises an air supply system for creating a protective atmosphere for sedimentation repairing, a power supply, a control box, an ultrasonic auxiliary spark sedimentation repairing and polishing integrated element, a workpiece chuck and a tool electrode, wherein the ultrasonic auxiliary spark sedimentation repairing and polishing integrated element comprises a shell, a spindle and an ultrasonic vibrator; the ultrasonic vibrator comprises a rear matching cover, two electrode plates of the ultrasonic vibrator, two piezoelectric ceramics, an insulation sleeve, an amplitude transformer, a pre-stress bolt and a electrode chuck. The device of the invention is portable and multifunctional, can repair the failure surfaces of module and parts through the ultrasonic spark sedimentation repairing, can improve the surface quality of the repaired sedimentation layer by ultrasonic polishing, can locally strengthen the surfaces of the module and parts through ultrasonic spark sedimentation and can remove surplus materials of the surfaces of the module and parts through reversed polarity spark processing. The portable ultrasonic auxiliary spark sedimentation repairing and polishing integrated processing method of the invention is simple in operation, is convenient and is practical.

The present invention relates to an image synthesis restoration method based on local structure features, comprising the following steps: inputting an image to be restored; initializing confidence coefficients of all pixel points in the image to be restored; selecting a template block of each pixel point by respectively taking each pixel point on a damaged boundary of the image to be restored as the center; calculating the priority of the template block where each pixel point on the damaged boundary of the image to be restored locates, and performing descending sort; searching an area which has the highest similarity with the template block having the greatest priority from an undamaged area in the image to be restored to serve as an optimal matching block; copying pixel values in the optimal matching block to the template block; and updating the confidence coefficients and returning to execute corresponding steps until the image to be restored is completely restored. The image synthesis restoration method based on local structure features provided in the present invention can improve image restoration accuracy for a damaged image with a complicated texture structure. The present invention also relates to an image synthesis restoration system based on local structure features.

The invention discloses a three-dimensional fine blood vessel reconstruction method. The three-dimensional fine blood vessel reconstruction method comprises the following steps of 1, loading a three-dimensional fine blood vessel original image; 2, performing blood vessel property-based image preprocessing; 3, performing blood vessel property-based characteristic matching; and 4, performing image filling and boundary processing. The invention furthermore discloses a three-dimensional fine blood vessel reconstruction system. With the adoption of the three-dimensional fine blood vessel reconstruction method and system provided by the invention, a filling region for repairing a missing part is obtained by mapping a position relationship and calculating an overall matching degree, and the missing part is repaired through the filling region to obtain a complete high-precision image, so that the accuracy of three-dimensional fine blood vessel reconstruction is high.

The invention discloses a batch repair method aiming at damaged plant leaves. Firstly, image data is pretreated. The pretreatment comprises the following steps: image reduction -grayscale transformation - Gaussian filter - binarization; Contour trimming: canny edge detection -finding edge contour- circumscribing minimum quadrilateral - ROI region trimming; Watershed segmentation: foreground marking-background marking- unknown region marking- watershed segmentation; Then, the image to be repaired is replaced by the same color, and the mask of the region to be repaired is made. Finally, the repaired part is repaired by FMM algorithm. Compared with the prior art, the invention can complete the incomplete part of the leaf, has high repair accuracy, and avoids the influence of the leaf incomplete caused by the force majeure on the subsequent experiment. The invention solves the problem of batch production of a mask through a series of image preprocessing operations, and realizes batch processing operation of damaged plant leaves.

The invention discloses a cutter abrasion repairing method. The cutter abrasion repairing method comprises the steps of firstly preprocessing an abraded cutter, clamping and positioning the preprocessed abraded cutter, enabling an image acquisition system to conduct data acquisition on the positioned abraded cutter and transmit acquired data to a computer so as to form an abraded cutter three-dimensional model, reading the cutter three-dimensional model into a computer before the cutter is abraded, calculating the cutter three-dimensional model required to be compensated through Boolean operation, filling a repair material in the cutter three-dimensional model required to be compensated, using laser direct-sintering technology to sinter and repair the abraded position, and finally cleaning and grinding the repaired cutter so as to complete cutter repair. The cutter abrasion repairing method can be applied to repair of high-cost and high-repair-value cutters including integral cutters, blades, drill tools, large milling cutters and the like, facilitate resource saving and reduction of production cost and achieve green manufacture.

The invention discloses a method for repairing a damaged machinery component based on three-dimensional digitization. The method comprises: firstly, a laser measurement system is adopted to obtain the superficial three-dimensional data points of a component to be measured; all the obtained superficial three-dimensional coordinate data information of the part to the measured are spliced; then the spliced three-dimensional coordinate data information is combined with a three-dimensional digitalized design method by characteristic recognition and extraction methods; the damaged part of the component to be measured is repaired, thus obtaining the three-dimensional physical model, the two-dimensional engineering drawing and the error analysis chart of the component to be measured; and then a technical document is returned to, thus duplicating and manufacturing the damaged part. The method has high repairing accuracy, the digitalized modeling is easy to edit and modify, the measurement cost is lower, the operability is strong, and the method is convenient for popularization and application.

The invention discloses a logic oil duct control system with electromagnetic switch valves and a failure autodiagnosis method. The logic oil duct control system comprises a plurality of electromagnetic switch valves, a speed regulator, one-way valves, a flow regulator, a hydraulic cylinder and an electric control unit, wherein a displacement sensor in the hydraulic cylinder, a piston cavity of the hydraulic cylinder and a pressure sensor connected to a rod cavity hydraulic circuit are respectively connected with the electric control unit through cables; the piston cavity of the hydraulic cylinder and the rod cavity hydraulic circuit are respectively connected with an oil port of each electromagnetic switch valve through the speed regulator; the piston cavity of the hydraulic cylinder and the rod cavity hydraulic circuit are respectively connected with the flow regulator and the oil port of each electromagnetic switch valve through the one-way valves; and a hydraulic circuit of a hydraulic cylinder rod cavity is connected with an oil return port through an overflow valve. The logic oil duct control system can accurately and automatically judge the failure of a certain or more electromagnetic switch valves, and the operating personnel can accurately replace the failed electromagnetic switch valve in time, thereby greatly saving the maintenance time, enhancing the maintenance efficiency and ensuring the reliable operation of equipment.

A method for determining a root cause of a problem in a multiple-element system includes correlating an incoming alarm signal with an element x in a multiple-element system and accessing an implication list comprising a list of all elements upstream of element x. At least one element on the implication list is weighted with data relating to the at least one element. Taking into account the weighting step, a probability is calculated that an element on the implication list comprises a most-probable root cause of the subsequent alarm signal. The identified most-probable root cause of the incoming alarm signal is output.

The embodiment of the invention discloses a method for repairing a driving route. The method comprises the following steps: a server reads a positioning data sequence of missing positioning data from a database storing positioning data, wherein the positioning data sequence includes a plurality of positioning data of the same moving object; the server plans driving routes for the missing positioning data in the positioning data sequence according to a path planning algorithm; the server uses the planed driving routes as candidate driving routes, and calculates a traveling time of each candidate driving route; and the server selects a driving route whose travelling time is the closest to the missing time of the positioning data as a missing driving route. Through the adoption of the method, the accuracy of repairing the missing driving route can be improved as much as possible, and the missing driving route that is as realistic as possible can be obtained. The embodiment of the invention further discloses a device for repairing the driving route.

The invention discloses an efficient image restoration method based on an alternating direction method. The efficient image restoration method based on the alternating direction method includes: translating image restoration problems into low rank matrix complement problems; further dividing each low rank matrix into a plurality of partitioned matrixes by decomposing an original image into three low rank matrixes according to R / G / B channels; complementing each partitioned matrix; combining all the partitioned matrixes after being complemented so as to obtain corresponding complement matrixes; further obtaining a repaired image according an combination of the complement matrixes. The efficient image restoration method based on the alternating direction method facilitates parallel processing through partitioning, complements each partitioned matrix by using the alternating direction method, greatly improves image restoration speed, can not damage overall relevance of the original matrixes due to the fact that the plurality of the partitioned matrixes are obtained by dividing the three low rank matrixes according to relative independence of the three low rank matrixes, and facilitates improvement of restoration accuracy. The efficient image restoration method based on the alternating direction method not only can be used for image restoration, but also can be used in fields of object removal, image noise reduction and the like.

The invention relates to a repairing device for an anode steel claw which is a consumptive part in the field of electrolytic aluminium, belonging to the electrolytic aluminium industry. The invention has the technical scheme that in a device for repairing an electrolytic aluminium anode steel claw by total cross-section fusion welding, a slag pool is formed by metal pallets and a cavity formed by moulds and a residual steel claw leg; the metal pallets are respectively arranged at the two opposite sides of the residual steel claw leg; the moulds are respectively positioned on the metal pallets and are closely jointed with the residual steel claw leg; the upper part and the lower part of each mould are respectively connected with a water outlet pipe and a water inlet pipe, and the interiors of the moulds are cavities; fluxing agent is arranged in the slag pool; an electrode is inserted into the slag pool, and a certain distance is maintained between the electrode and the residual steel claw leg; and the electrode and the residual steel claw leg are connected with a transformer by leads. The steel claw leg repaired by the device is high in strength and utilization rate; and meanwhile, the energy is saved, the consumption is reduced, and the problems caused by artificial repairing can be solved.

The invention discloses an image composition beautifying method based on a visual area proportion balance rule. The method comprises the following steps: (1) inputting an image; (2) detecting an area segmentation line of the image and dividing the image into two areas; (3) obtaining an area to be optimized and an optimization strategy of the image by utilizing the visual area proportion balance rule and based on the area segmentation line; (4) calculating texture complexity of the area to be optimized; (5) performing optimization on the area to be optimized based on the optimization strategy, utilizing a rapid seam-insertion method to perform stretching or reducing on the area to be optimized when the texture complexity of the image is smaller than a preset threshold value of the texture complexity, and otherwise, utilizing a sample-based high-efficiency image restoration algorithm to perform stretching on the area to be optimized; and (6) outputting the optimized image. The invention further discloses an image composition beautifying system based on the visual area proportion balance rule. According to the image composition beautifying method and system based on the visual area proportion balance rule, the image composition is automatically beautified and the image is allowed to have a better esthetic visual effect.

The invention discloses a thin-wall component recombining and restoring method based on pulse welding and laser augmented material restoring. The method is characterized in that firstly, the distribution and trend of cracks and other penetrating damage are determined through a coloring check method; secondly, grinding treatment is conducted through a hard alloy rotary file, and a penetrating groove in a relatively regular shape is formed; thirdly, optimal grinding of a pulse welding groove is conducted; fourthly, cleaning of the penetrating groove and a groove body of the penetrating groove is conducted; fifthly, pulse welding bridging restoring of the penetrating groove is conducted; sixthly, a bridging welding restored body is ground and cleaned; and seventhly, laser augmented material restoring is conducted on a damaged structure. The method has the following advantages that the pulse welding restoring bridging capability is high, a base plate does not need to be additionally designed or manufactured, the restoring technology process of a penetrating damage component can be simplified, and the restoring efficiency is improved. A pulse welding restoring control system outputs data stably and accurately, heat input in the restoring process can be accurately controlled, the deformation of the restored workpiece can be reduced, and the restoring precision is high. Three-dimensional forming restoring of the damaged component can be automatically achieved through laser deposition restoring, the restoring efficiency is improved, and the stability and consistency of the restoring technology are good.

The invention relates to a wall painting machine based on the internet of things and with a repairing function and good polishing effect. The wall painting machine comprises a base and four moving wheels, and further comprises a repairing mechanism, a storage chamber, a connecting plate, two lifting mechanisms and four polishing mechanisms. According to the wall painting machine based on the internet of things and with the repairing function and good polishing effect, lifting of the four polishing mechanisms is driven through the lifting mechanism, a wall surface to be painted is polished, thus, the wall surface is smooth and neat, a moving assembly drives the polishing mechanisms to move in the front-back direction, thus, the polishing mechanisms polishes the wall surface from top to bottom, the circumstance that while the polishing mechanisms polish the wall surface in the horizontal direction, the wall surface is polished in the vertical direction, and thus the wall surface is coarse is avoided, the polishing effect of the wall painting machine is good, the lifting mechanism is stable, recesses and cracks of the wall surface can be repaired through the repairing mechanism, the integral attractiveness of the wall surface is improved, and therefore, the practicality of the wall painting machine is high.

The invention relates to a bidirectional detection and restoration method for traffic flow abnormal data based on a KNN algorithm, and the method comprises the following steps: 1), obtaining normal traffic flow historical data, taking each five pieces of continuous data as a group, and building a historical data state vector library; 2), obtaining an abnormal value in a to-be-restored traffic flowdata, and marking the abnormal value as v(w); 3), constructing an abnormal data state vector X according to the abnormal value; 4), calculating the Euclidean distances d between the abnormal data state vector X and all historical data state vectors Xn, carrying out the optimization, and obtaining k groups of historical data state vectors after optimization and the corresponding Euclidean distances di (i=1, 2, ...k) 5), calculating a restoration value v(w)' according to the k groups of historical data state vectors after optimization and the corresponding Euclidean distances di; 6), carrying out the deletion, filling and restoration of the abnormal value. Compared with the prior art, the method is high in restoration precision, is high in applicability, improves the restoration precision,and improves the traffic data quality.

The invention provides an improved Criminisi image restoration method with self-adaptive gradient block division and equilong transformation. The method is characterized by, to begin with, marking known and restoration regions of an image to be restored, calculating a gradient histogram for all pixel points in the known region, and dividing the pixel points in the known region into smooth, texture and edge three types of pixels; then, enhancing discrimination capability for an edge texture portion through a block-classification improved priority function to solve the problem of texture extension caused by a conventional method; next, judging the block type of blocks to be restored by judging known gradient information in the block to be restored, and establishing a self-adaptive block size function to ensure that different types of blocks to be restored are restored by different sizes of blocks respectively, and carrying out matching only in the corresponding types to improve matching efficiency; and finally, introducing the equilong transformation to improve matching precision. Compared with the existing method, the method well overcomes the problems of texture extension and high time complexity and the like, and the restoration quality is improved.

The invention discloses an image restoration method and device based on a Shannon-Nuttall wavelet multi-scale expression. The method comprises the following steps: obtaining a to-be-restored image; shielding a to-be-restored area in the to-be-restored image; constructing a multi-scale interpolation operator based on a Shannon-Nuttall interpolating wavelet and realizing the geometric description of the to-be-restored image under an interpolating wavelet framework; constructing an image restoration variational model under the interpolating wavelet framework according to the geometric description of the to-be-restored image under the interpolating wavelet framework; solving the image restoration variational model under the interpolating wavelet framework and obtaining a sharp image. According to the image restoration method and device disclosed by the invention, the image is reconstructed through the multi-scale interpolation operator, the image restoration variational model is constructed under the interpolating wavelet framework, and the image is solved through a sparse grid algorithm to obtain the restored image, so that the efficiency and the precision of image restoration are improved.

The invention relates to the technical field of computer software, and discloses a method for automatically repairing a program crash defect based on question-answer website analysis. The method comprises the steps of obtaining multiple related web pages according to a crash track when a program crashes, wherein each related web page comprises a defect code for describing the crash and a repairing code for repairing the crash; extracting a fragment of the defect code and a fragment of the repairing code in each related web page; generating an edit script for repairing the crash according to the fragment of the defect code and the fragment of the repairing code; determining the position of an error code in a program source code according to the crash track and the defect code, and applying the edit script to repair the error code. According to the method disclosed by the invention, a repairing scheme can be automatically searched on the related web pages according to the crash track, and the corresponding edit script can be generated for repairing the crash, so as to get rid of the dependency on an operator during a repairing process, automatic repairing is realized, and the repairing accuracy is higher.

The invention discloses a method for finely restoring karst paleogeomorphology in the technical field of paleogeomorphology restoration. The specific steps of the method for finely restoring karst paleogeomorphology are as follows: S1: Find a relatively accurate representative paleosea in the overlying stratum of ancient weathering crust Plane stratigraphic depositional interface; S2: Obtain the relative height of the topography on the weathering crust through the stratum thickness between the depositional interface and the weathering crust; S3: Use the method of high-resolution sequence stratigraphy to calculate the relative height of the topography in step S2 S4: Construct the seismic profile; S5: Analyze the seismic profile features on the seismic profile, and restore the karst ancient landform. The restoration image of the karst ancient landform in the present invention is intuitive and realistic, with low cost and easy Realized that the restoration of ancient landforms is relatively complete, and at the same time, the impact of the later structure of the landforms is considered, and the restoration accuracy is high.

The invention discloses a data repairing method for motion capture. The method comprises the following steps: acquiring two or more than two frames of motion image data of a plurality of marked points on a moving object and respectively tracking the motion tracks of the plurality of marked points so as to determine the missing points and efficient points of the frames; computing the parameter sets of two or more than two efficient points adjacent to the missing point in the previous frame and the sum of the distances between the missing point and the two or more than two efficient points; andrespectively computing the weights of the motion vectors formed by the two or more than two efficient points and the points corresponding to the two or more than two efficient points in the current frame according to the parameter sets and the computed sum of the distances between the missing point and the two or more than two efficient points and obtaining the position of the missing point in the current frame by combining the computed weights and the parameter sets. When the data repairing method is applied to motion capture of the facial expressions, motion mechanisms of facial muscles which belong to a nonrigid structure are thoroughly considered, thus improving the repairing precision of the missing points.

The invention belongs to the technical field of denture processing, and relates to a three-dimensional printing preparation method for the upper dental crown of a dental implant. The method specifically comprises the six steps of dental model scanning, scanning data processing, dental crown restoration design, printing powder preparation, typesetting data transmission and dental crown restorationprinting, wherein printing powder is prepared from albite, potassium feldspar, quartz, talcum powder, borax, sodium carbonate, potassium carbonate, sodium borate and cerium dioxide, wherein the printing powder is nano-scale fine particles. The three-dimensional printing preparation method for the upper dental crown of the dental implant has the advantages that the printing powder is prepared to control the color and transparency of a restoration, a computer controls a three-dimensional printer for molding through one-time printing, cumbersome procedures of traditional processing are eliminated, the work efficiency is high, the repair accuracy is high, a molded product has good compatibility with the oral cavity, the defect rate is low, and defective products are prevented from causing waste of raw materials; the production method is scientific in principle, easy to implement, safe and reliable, and the method is friendly to the application environment.

The invention relates to the technical field of artificial intelligence, and provides a face image restoration method and device, a face image restoration model, a medium and equipment. The method comprises the following steps: acquiring a to-be-restored image, and acquiring a face mark point image corresponding to the to-be-restored image; inputting the to-be-restored image and the face mark point image into a restoration model of the face image; carrying out at least two times of spatial feature conversion processing on the to-be-restored image and the face mark point image based on a restoration model of the face image to obtain fused face features; and performing convolution processing on the fused face features, and determining the output of the restoration model of the face image as a restored image corresponding to the to-be-restored image. According to the technical scheme, the face restoration accuracy can be improved.