1980 results about "Image signal" patented technology

The invention relates to an underwater robot navigation positioning method and system. The method comprises the following steps: S1, acquiring angular velocity and acceleration signals through an inertial measurement unit; s2, resolving a three-dimensional velocity observation value according to the beam radial velocity vector signal in combination with the angular velocity signal, and extracting environment feature point cloud data according to the acoustic image signal; s3, multi-source data time synchronization is carried out, and a fusion input signal with time-space alignment is generated; s4, constructing an adaptive factor graph optimization model, and dynamically adjusting an inertial navigation solution node based on a real-time weight coefficient; inputting the environment feature point cloud data into a closed-loop detection module to generate a loopback factor node, and adaptively correcting the weight of the node according to the feature matching degree; and S5, solving the adaptive factor graph optimization model through a nonlinear optimization algorithm. According to the underwater robot navigation positioning method and system, the problem that the fusion positioning precision of a multi-source heterogeneous sensor is insufficient in an underwater GPS-free environment can be solved.

The present invention relates to the technical field of cross-modal image signal reconstruction. Disclosed is a cross-modal joint source-channel coding and decoding method adaptable to changeable scenarios. The method comprises: first designing a Transformer encoder-based cross-modal channel coding and decoding optimization solution, so as to achieve the performance improvement and robustness of a channel encoder and a channel decoder; then designing a cross-modal source coding and decoding optimization solution for haptic-to-image generation based on a latent diffusion model, so that under an image signal loss scenario, haptic information is used to guide image generation; and finally, incorporating transfer learning technology, so as to reduce additional training costs caused by a system facing changeable cross-modal communication scenarios such as a changeable channel signal-to-noise ratio and different transmission tasks. Under cross-modal changeable communication scenarios, the joint source-channel coding and decoding method provided in the present invention can solve the problems of the inability of a receiving end to well complete image reconstruction, and additional model training costs caused by changeable channel environments and scenarios.

The invention belongs to the field of machining, and relates to a machining process quality online monitoring and feedback adjusting system and method. The system comprises a multi-modal data acquisition module used for acquiring a mechanical vibration signal and a high-frequency micro-damage signal; the feature preprocessing and cross-modal fusion module is used for carrying out time domain, frequency domain and time-frequency domain feature extraction on the collected mechanical vibration signals and high-frequency micro-damage signals, carrying out edge, texture and shape feature extraction on workpiece machining surface image signals, and introducing a cross-modal attention mechanism to achieve weighted fusion of heterogeneous signals. And finally generating a fusion feature vector with high discrimination. The machine tool / cutter state and workpiece quality intelligent analysis module is used for constructing an intelligent analysis model for the equipment operation state and the workpiece machining quality; inputting the fusion feature vector into an intelligent analysis model to obtain a quality evaluation result; and the quality judgment and feedback adjustment module is used for adopting a dynamic threshold adjustment mechanism to realize adaptive comparison with a process standard according to a quality evaluation result.

The invention relates to the field of image data processing, and discloses a texture preserving type image denoising and enhancing method based on a generative adversarial network, which comprises the following steps: acquiring an original image signal, and calculating low-frequency sub-band data and high-frequency sub-band data by using discrete wavelet transform; calculating the gradient magnitude of the low-frequency sub-band data to generate a structural significance gradient map; establishing a reverse mapping relation based on the structure saliency gradient map, and generating a spatial self-adaptive dynamic gating threshold; performing statistical gating on the high-frequency sub-band data by using the dynamic gating threshold to generate a high-pass gain coefficient and a low-pass suppression coefficient; according to the method, cross-band modulation logic of the structure flow to the texture flow is established, so that the problem that weak texture signals are easy to lose under non-uniform illumination is solved, and non-structured noise filtering and structured micro texture restoration are realized on the premise of not depending on semantic tags.

A system and method are disclosed for low-light image enhancement using hierarchical adaptive wavelet decomposition with cross-scale feature fusion. The system analyzes a raw input image to determine image characteristics and preprocessing parameters. A hierarchical adaptive wavelet decomposition process creates a variable-depth decomposition tree comprising frequency domain nodes, with decomposition depth determined by local image complexity. Cross-scale feature fusion implements attention mechanisms between nodes at different decomposition levels, enabling bidirectional information flow across scales. A dynamic network pool allocates specialized neural networks to process nodes based on their frequency characteristics, with weight sharing between similar nodes for efficiency. An adaptive reconstruction engine traverses the decomposition tree using learned filters and multi-scale residual learning to produce an enhanced image. The hierarchical approach enables superior low-light image enhancement by allocating computational resources based on content complexity, achieving better quality than fixed decomposition methods while maintaining compatibility with existing image signal processing pipelines.

The invention provides an offshore wind turbine generator fault diagnosis method and system based on multi-source sensor data fusion. The method comprises the steps that a vibration signal from at least one component of a wind turbine generator is acquired through a vibration sensor; performing time-frequency conversion on the vibration signal by applying synchronous compression wavelet transform to obtain time-frequency representation of the vibration signal; when the reconstruction error exceeds a preset threshold value, it is judged that an abnormal event exists in the vibration signal; obtaining the position of a part corresponding to the abnormal event; starting an image sensor and an acoustic sensor according to the position of the component, and acquiring an image signal and a sound signal of the component according to the image sensor and the acoustic sensor; according to the DS evidence theory, the vibration signal, the image signal and the sound signal, obtaining the confidence of the fault type; the fault type of the component is judged according to the maximum confidence allocation principle, high-resolution time-frequency analysis can be achieved through synchronous compression wavelet transform (SST), and the fault feature identification degree is improved in combination with the self-encoding neural network and the D-S evidence theory.

The invention belongs to the field of computer vision image signal processing, and relates to a multi-robot collaborative indoor scene semantic segmentation method and system based on multivariate interactive learning. The method comprises the steps of adopting an image feature encoder based on a double attention integration module to realize cross-modal feature fusion on an RGB image and a depth image; a dimension anisotropy attention enhancement module is adopted, and feature optimization is carried out through a multi-dimensional direction attention mechanism; a cross-task collaborative interaction module is adopted, and information is shared among different tasks by utilizing an interaction mechanism based on a feature level; in a multi-task heterogeneous decoding stage, each robot uses a decoder with a specific task to execute a specific task, and semantic segmentation, instance segmentation and scene classification are realized; and a joint optimization loss function is adopted to perform end-to-end joint training on multiple tasks, so that overall optimization and collaborative awareness performance improvement are realized. According to the method, the precision and robustness can be improved when complex indoor scene semantic segmentation and other visual tasks are executed.

The invention relates to an unmanned aerial vehicle inspection method and system based on an overhead transmission line, and the method comprises the steps: S1, generating an unmanned aerial vehicle three-dimensional space coordinate signal, and outputting a position correction signal based on a tower topology database; s2, in response to the position correction signal, controlling a multispectral sensing module to synchronously acquire a visible light image signal, an infrared temperature signal and a laser radar point cloud signal of the line; s3, inputting the visible light image signal into a defect identification model to generate a first defect feature signal; s4, fusing the first defect characteristic signal and the second defect characteristic signal to form a cross-modal diagnosis signal, and triggering a defect classification decision signal; and S5, classifying decision signals and cross-modal diagnosis signals according to defects. According to the air quality intelligent monitoring method and system based on sensing data feedback, the problems of inaccuracy in routing inspection positioning, defect missing detection and insufficient endurance in a complex environment can be solved.

The present application provides a battery thermal runaway detection method and apparatus. The method comprises: acquiring a time series signal generated when a gas sensor identifies a gas component released by a battery, and converting the time series signal into an image signal, so as to strength key features in the signal; processing the image signal by means of a target detection model, so as to predict a gas type and a gas concentration; and then determining a thermal runaway risk of the battery on the basis of a prediction result. Therefore, by introducing a machine learning algorithm, simple gas identification requirements are met, and complex data analysis and prediction can be carried out, thereby accurately predicting the thermal runaway risk of lithium batteries, and effectively improving the early warning precision of thermal runaway.

The invention relates to the technical field of power transmission line detection, and provides a high-voltage overhead power transmission line insulator string detection device and a control method thereof. Comprising an unmanned aerial vehicle body, and an airborne data transmission module, an airborne insulator detection module and an image acquisition and analysis module which are carried on the unmanned aerial vehicle body, the image acquisition and analysis module is used for acquiring an image signal of the insulator string and analyzing the image signal so as to output a position control signal of the unmanned aerial vehicle and an angle control signal of the airborne insulator detection module; the unmanned aerial vehicle body flies to a specified position based on the position control signal; the airborne insulator detection module is used for rotating to a specified angle based on the angle control signal and detecting an insulator string; and the airborne data transmission module is used for receiving the detection data sent by the airborne insulator detection module and determining the health state of the insulator chain based on the detection data. Zero-value insulator live-line detection of a full-voltage-class full-string type insulator string of a power transmission line can be realized.

The invention discloses an underwater image de-scattering method, system and device for scattering field decoupling in the technical field of underwater optical imaging and computer vision, and the method comprises the steps: calculating an intensity graph and a scene linear polarization degree graph according to an obtained multi-scale underwater image; according to the intensity image, performing image signal decomposition by using a dynamically constrained background light separation model to obtain a background scattering component and a target feature component; according to the scene linear polarization degree map and the background scattering component, calculating fusion transmissivity according to a non-linear optical transmission equation; carrying out noise suppression processing on the fusion transmissivity by adopting a double-constraint optimization architecture to obtain the fusion transmissivity subjected to noise suppression processing; and performing image reconstruction through a reverse optical propagation model according to the fusion transmissivity and the target feature component subjected to noise suppression processing to obtain a scattering-removed underwater image. The method effectively breaks through the dependence of a traditional method on the uniform hypothesis of a scattered field, and provides high-precision visual restoration support for an underwater detection task.

The invention discloses an LED liquid crystal display screen brightness and color difference detection and correction method and system, and belongs to the technical field of liquid crystal display. Initial brightness values and color values of all display units of a display screen are obtained; constructing a brightness and color difference reference model based on the convolutional neural network, and calculating a brightness deviation delta L and a color deviation delta C between the brightness and color difference reference model and a standard value; constructing a brightness-color difference error matrix W; according to the error matrix and in combination with a self-adaptive threshold algorithm, identifying a brightness-color difference abnormal region and corresponding coordinates of the brightness-color difference abnormal region; based on a deviation scoring result, introducing a gradient regression model to generate a brightness gain coefficient and a color adjustment coefficient, and constructing a correction parameter set P; the parameter set is issued to a control chip, dynamic image signal modulation and local frame level refreshing are executed, and point-by-point brightness and color difference correction is achieved; the method has the advantages of high precision, high automation, high adaptive capacity and the like, and is suitable for improving the brightness and color difference uniformity in various LED liquid crystal display screen scenes.

A camera module includes an imaging lens having an optical axis, a variable through hole mechanism, an image sensor, an image sensor actuating module and a flexible printed circuit. The variable through hole mechanism is disposed corresponding to the imaging lens to adjust an amount of light into the imaging lens. The image sensor is disposed on the optical axis and is configured to convert imaging light passing through the imaging lens into an image signal. The image sensor actuating module is configured to move the image sensor. The variable through hole mechanism and the image sensor actuating module each have a coil disposed on the flexible printed circuit. The flexible printed circuit is electrically connected to the variable through hole mechanism and the image sensor actuating module.

The invention relates to a naked eye 3D binocular image acquisition and real-time processing system based on hardware synchronous triggering, and belongs to the technical field of image processing and three-dimensional display. According to the system, a hardware synchronous triggering mechanism is adopted, a time sequence synchronous control unit is used for sending a synchronous pulse signal to a binocular image sensor, and strict synchronization of left and right viewpoint image acquisition is ensured. The image signal processing unit processes collected original data, the stereo parallax correction module performs epipolar correction, and the sub-pixel interleaving module generates a composite view frame according to grating parameters of the display terminal. The system realizes high-speed data transmission through a double-buffer direct memory access DMA mechanism. The problems of visual tearing and weak stereoscopic impression caused by asynchronous binocular image acquisition in the prior art are solved, nanosecond-level synchronization precision is realized, optical crosstalk is reduced, and smoothness and comfort of stereoscopic display are ensured.

The invention discloses a farmland soil moisture content real-time surveying system and method based on unmanned aerial vehicle multispectral imaging. The method comprises the steps that multispectral sensor parameters are configured, and an unmanned aerial vehicle flight scanning path is planned; a space coordinate system and time synchronization mechanism is established, and high-precision positioning and time synchronization are achieved; collecting and calibrating visible light, near-infrared and short-wave infrared original image data to form a spectral signal set; calculating an image signal-to-noise ratio, screening a low-quality image, and executing noise suppression; extracting timestamp information from the suppressed image, performing spatial positioning, and establishing a spatial corresponding relation database; calculating time synchronization precision and performing image correction; evaluating the spatial resolution according to the distortion coefficient; calculating a comprehensive quality parameter, comparing the comprehensive quality parameter with a threshold value, and obtaining a soil moisture content distribution parameter when a condition is met; a farmland moisture state map is generated, partition marking is carried out according to a multi-level standard, a monitoring result is output, and high-precision and real-time monitoring of the farmland soil moisture content is achieved.

The invention discloses an ultra-wide-spectrum endoscopic imaging device capable of being used for short-wave infrared fluorescence imaging, which is used for in-vivo minimally invasive imaging and comprises an illumination module, a wide-spectrum image transmission module, an imaging module and an image processing module. The illumination module consists of a light source and a light guide optical fiber, and is used for generating illumination light from ultraviolet to short wave infrared band and transmitting the illumination light to the wide-spectrum image transmission module; the wide-spectrum image transmission module consists of a few-interface optical imaging channel and a light guide illumination channel, and is used for illuminating an imaging target, exciting target fluorescence and collecting image signals from ultraviolet to short-wave infrared bands; the imaging module is composed of an optical filter and an imaging camera, and the imaging camera can be a wide spectral response image sensor or can be formed by combining cameras working in different wavebands and is used for converting optical images in different wavebands into image data; and the image processing module is used for registering and fusing image data of different wavebands in real time and displaying the image data.

The invention relates to the field of textile detection. The invention discloses a detection system for textile defect detection, which comprises a light source arranged on one side of a textile and used for outputting a detection light beam; the light path adjusting assembly is located in a light path from the light source to the textile; the image acquisition device is arranged on the other side and is used for acquiring an irradiated image signal; the angle adjusting device adjusts irradiation and acquisition angles of the light source and the image acquisition device; a reflecting component is arranged on one side of the light source and is controlled by a control switch to reflect a detection light beam to the textile; and the image processing unit is connected with the image acquisition device, analyzes the image signal and outputs a defect detection result. According to the detection system, light and shade contrast of flaw edges is enhanced through double-beam interference, and fine defects are accurately captured in combination with angle adjustment and multi-light-source adaptation. The detection method adjusts parameters in stages, automatically matches the optimal detection angle and light source intensity, and improves the adaptability of complex fabric detection.

The invention relates to a high-signal-to-noise-ratio ultrasonic imaging method and device for internal defects of a curved surface component, a medium and a product, and belongs to the technical field of ultrasonic imaging. According to the method, a linear array ultrasonic probe is adopted to carry out single-time ultrasonic excitation on a curved surface component, a full-matrix echo signal is obtained, first arrival time extraction is carried out on the collected signal, an ultrasonic incident angle is estimated, and discrete coordinates of the surface of the component are obtained; the method comprises the following steps of: fitting a curved surface model, constructing an equivalent virtual array model, calculating transmitting and receiving delay time of each pixel point, realizing initial full-focusing beam forming of signals, estimating an expected signal variance and a noise variance based on maximum likelihood estimation, constructing an initial coherent beam forming machine and performing iterative updating, and finally forming a beam. And taking the coherent beam synthesizer obtained by iteration as a weighting factor, acting on an initial beam synthesis signal for weighting, and completing image reconstruction. The method has the advantages of being high in calculation efficiency, low in system overhead, high in imaging signal-to-noise ratio and the like, and accurate imaging can be achieved in a curved surface structure scene.

The invention provides an intelligent wind power mixed tower damage detection method and system based on multi-modal data fusion. The method comprises the following steps: acquiring a vibration signal, an image signal and an acoustic emission signal of a wind power mixed tower through a multi-modal data acquisition layer; performing timestamp synchronization and space coordinate mapping alignment on the vibration signal, the image signal and the acoustic emission signal by adopting a bidirectional LSTM network to obtain multi-modal heterogeneous alignment data; based on multi-scale feature extraction, identifying vibration features of a steel tower section and image features and acoustic emission features of a concrete section in the multi-modal heterogeneous alignment data; dynamically distributing feature weights for the vibration features, the image features and the acoustic emission features by using a cross-modal attention network, and determining the damage probability and the expansion speed of the wind power hybrid tower; based on the damage probability and the expansion speed, graded early warning is executed, and the wind power mixed tower damage detection efficiency is effectively improved.

The embodiment of the invention discloses a scanning NV probe microscope which comprises an atomic force microscopy system used for detecting surface topography information of a sample and an optical detection magnetic resonance system used for detecting surface magnetic field distribution information of the sample, and the atomic force microscopy system comprises a vibration arm and a diamond probe. The optical detection magnetic resonance system comprises an NV color center, a transition excitation light path and a fluorescence detection light path; a sample to be scanned and the transition excitation light path are respectively positioned on two opposite sides of the diamond probe, and the transition excitation light path and the fluorescence detection light path are positioned on the same side of the diamond probe; a first metal coating film is arranged on one side, close to the sample, of the diamond probe and is used for blocking surface fluorescence of the sample. According to the scanning NV probe microscope provided by the embodiment of the invention, the metal coating is arranged at the bottom of the diamond probe to block stray fluorescence from the surface of the sample, so that the problems of low magnetic field imaging signal-to-noise ratio and limited measurement precision of the existing scanning NV probe microscope are solved.

A high-order driving auxiliary lane changing trajectory planning method for an intelligent vehicle comprises the steps that laser radar and camera data are obtained, and feature vectors of obstacles are extracted after the data are processed by an ISP at the front end of a chip of the intelligent vehicle; constructing a spatio-temporal joint attention network model, inputting the feature vector of the obstacle into the spatio-temporal joint attention network model, calculating an interaction weight between the vehicle and the obstacle through a self-attention mechanism, and synchronously outputting predicted trajectory distribution of the obstacle and a candidate planning trajectory cluster of the vehicle; introducing a dynamic game in the game theory, and constructing a comprehensive cost function in combination with the interaction weight to obtain a screened track; and taking the screened trajectory as an initial value, inputting the initial value into a rear-end quadratic programming optimizer on the premise of meeting vehicle dynamics constraints, carrying out numerical optimization solution, searching an optimal programming sequence, and finally finishing lane changing trajectory planning. According to the invention, the interactive perception capability and riding comfort of the intelligent vehicle in a complex traffic scene can be improved.

An endoscopic imaging system for use in a light deficient environment includes an imaging device having a tube, one or more image sensors, and a lens assembly including at least one optical elements that corresponds to the one or more image sensors. The endoscopic system includes a display for a user to visualize a scene and an image signal processing controller. The endoscopic system includes a light engine having an illumination source generating one or more pulses of electromagnetic radiation and a lumen transmitting one or more pulses of electromagnetic radiation to a distal tip of an endoscope.

The invention discloses an endoscope image processing method based on oxyhemoglobin saturation and multiple central wavelengths, and relates to the field of medical image processing.The method comprises the steps that a first initial image signal, a second initial image signal, a third initial image signal and a fourth initial image signal are obtained; calculating oxyhemoglobin concentration information of the target part according to the second initial image signal and the third initial image signal; calculating deoxidized hemoglobin concentration information of the target part according to the second initial image signal and the fourth initial image signal; calculating the oxyhemoglobin saturation of the target part according to the oxyhemoglobin concentration information and the deoxidized hemoglobin concentration information; performing brightness enhancement on the second initial image signal based on the oxyhemoglobin saturation to obtain a second enhanced image signal; and generating a target image according to the second enhanced image signal, the first initial image signal and the third initial image signal, wherein the target image is used for identifying the bleeding position of the target part.

The invention discloses a liquid crystal display device for improving contrast based on movable light guide lattice points and a light guide layer and a control method of the liquid crystal display device. The device comprises a backlight module, a dynamic light guide layer, a liquid crystal display panel and a control system. The dynamic light guide layer is packaged by two transparent substrates to form a sealed cavity, the cavity is filled with transparent base liquid and movable micro units with different refractive indexes, and transparent electrode arrays are arranged on the inner surfaces of the substrates. The control system controls the electrode array to generate a corresponding electric field according to brightness and darkness distribution of an input image signal, drives the micro-units to move in the base liquid and gather to the position below an image brightness field area so as to enhance light output, and meanwhile enables the micro-units to be dispersed below a dark field area so as to reduce light leakage. According to the invention, a static light guide structure is changed into a dynamic programming system, and the real-time pixel-level matching of backlight distribution and display content is realized, so that the contrast ratio is remarkably improved, the halo is inhibited, and the power consumption is reduced. The invention also protects the core component of the dynamic light guide layer and the corresponding control method.

A display device includes a pixel unit that includes a plurality of pixels, a compensator configured to receive sensing data from the pixel unit in a sensing period and that detects and compensates for a characteristic of the pixels, and a timing controller configured to convert an image signal received from an outside into image data.

The invention relates to the technical field of smart home control, and particularly discloses a smart home central control system and method based on multi-mode perception, and the system comprises the steps: synchronously collecting a voice audio signal, a gesture image signal, an infrared thermal imaging signal and a millimeter wave radar signal through a plurality of groups of sensors; performing blind source separation processing on the voice and gesture signals, extracting a voice command component and a gesture action component which are independent in statistics, and performing space-time alignment and Kalman filtering fusion on the infrared and radar signals to generate a dynamic environment sensing graph; voice intention features, gesture track features and environment anomaly features are extracted through Mel frequency cepstrum coefficient analysis, skeleton key point tracking and multi-level convolution processing; constructing a three-dimensional decision matrix based on the features, performing weighted evaluation through a fuzzy logic rule base to generate a control instruction priority sequence, and dynamically adjusting an equipment operation mode according to the priority; according to the invention, the problems of control conflict and response delay caused by multi-mode signal coupling are solved.

The invention belongs to the field of image processing, and provides a projection image automatic correction method and system based on deep learning, and the method comprises the steps: controlling a projector to output a preset test image; acquiring a first image of a user at a viewing angle by the mobile terminal; acquiring a second image at a projection angle by a shooting device of the projector; extracting features and predicting adjustment parameters for projection correction based on a projection image correction model of deep learning; the adjustment parameters are issued to the projector through a network; and performing pre-distortion processing on an image signal to be projected based on the geometric mapping parameter, and performing pixel-level adjustment of brightness and color based on the compensation parameter to generate a corrected projection output image. According to the method, the usability and accuracy of projection image correction can be improved.

The invention belongs to the technical field of target detection, and particularly discloses a two-dimensional adaptive constant false alarm detection method for a gridding complementary reference sliding window, which comprises the following steps: constructing the gridding complementary reference sliding window under a detection matrix and distributing reference units; key parameters of a detection matrix are initialized, and an original image signal matrix is converted into a signal power matrix; based on the window-type topological structure, coordinate features of the reference unit are discriminated and assigned, and after traversal, a mean value is calculated to extract background clutters; and calculating a second-order statistic VI and a mean ratio MR, and selecting a constant false alarm algorithm to calculate a detection threshold through three threshold factor calculation formulas in combination with a judgment rule. According to the two-dimensional self-adaptive constant false alarm detection method of the gridding complementary reference sliding window, the detection accuracy is improved by 49.4% under the multi-target and clutter interference environment, the threshold stability and the boundary processing capacity are better under different backgrounds, the false alarm and missing report rate is effectively reduced, and the target detection performance is improved.

The invention provides a polarization hyperspectral water quality monitoring system and method, and the system comprises a tower footing polarization hyperspectral imaging module which is disposed on a building beside a to-be-detected water area, and carries out the area array staring scanning of the to-be-detected water area through a polarization hyperspectral imager, so as to obtain a water body polarization hyperspectral image signal of the to-be-detected water area; the image spectral data processing module is used for deducting dark current and spurious signals from the water body polarization hyperspectral image signals, and calculating the water body spectral reflectivity of the water area to be measured according to the water body polarization hyperspectral image signals by combining a gray value and reflectivity relation model constructed by using the hyperspectral image signals of the gray scale calibration plate; and the water quality parameter spatial distribution inversion module is used for performing inversion on the to-be-detected water quality parameters of the to-be-detected water area according to a pre-constructed water quality parameter inversion model and the water body spectral reflectivity. According to the technical scheme, the integration level is high, the automation degree is high, rapid deployment can be achieved, and all-day continuous unattended monitoring can be achieved.

The present application provides an information processing device and method, and a computer-readable storage medium. The information processing device comprises a processing circuit which is configured to: generate a saliency map of a sample image on the basis of a predetermined model which processes a task for the sample image, wherein the saliency map reflects the degree of attention to objects at different positions in the sample image when the predetermined model processes the task; and adjust, on the basis of the saliency map and the labeling area in the sample image, the parameters of the image signal processor which generates the sample image, such that the difference between the task processing result and the labeling value of the sample image meets a preset condition.