35results about How to "Does not involve complex calculations" patented technology

The invention relates to a driver fatigue detection method based on face video analysis, belonging to the field of vehicle active safety and mode recognition. The method comprises the steps of: carrying out face detection on a face image to be detected for primarily positioning local organs, such as eyes, nose and mouth; further obtaining face characteristic points positioned accurately; and finally, on the basis of an accurate positioning result of multiple frames of face characteristic points to be detected, carrying out quantization description on a face motion characteristic, and obtaining the result of fatigue detection on a driver to be detected according to a face motion statistical indicator. The driver fatigue detection method has high reliability, has no special requirements on hardware, does not involve complex operation, and has good instantaneity so as to be applied in real time in practice. The driver fatigue detection method has better robustness on the aspects of eyeglasses wearing, face angles, nonuniform illumination and the like, and can be suitable for various types of vehicles for reminding a driver of being in a fatigue state so that traffic accidents are avoided.

A redundancy control method based on double position sensors comprises the steps of using a redundancy decision module to carry out the extreme value fault discrimination on the output values of the two position sensors of a servo system, selecting to output the output value of the position sensor not generating an extreme value fault, if the two position sensors do not generate the extreme value faults, discriminating the consistency of two feedbacks, if the two feedbacks are consistent, keeping the constant output of the current position sensor, if the two feedbacks are inconsistent, carrying out the zero value fault discrimination, selecting to output the output value of the position sensor not generating a zero value fault, if the two position sensors do not generate the zero value faults, carrying out the feedback and instruction error discrimination, and outputting the output value of the position sensor having a small absolute value of an error value of the output value and an instruction. The redundancy control method based on double position sensors of the present invention is easy to realize, and is good in fault compatibility and high in reliability.

The invention provides a complete-polarization sea clutter simulation method. The complete-polarization sea clutter simulation method comprises the following steps: simulating a random sequence with K distribution by an SIRP (Spherical Invariable Random Process) method and utilizing the random sequence with K distribution as the sea clutter of an HH (Horizontal-Horizontal) polarization channel; then simulating the sea clutter of a VV (Vertical-Vertical) polarization channel by utilizing the amplitude constraint and phase constraint relation between the HH polarization channel and the VV polarization channel; and finally simulating the sea clutter of a cross polarization channel by utilizing the constraint relation between the HH polarization channel and the cross polarization channel, so as to obtain the complete polarization scattering matrix of each scattering unit, so that the simulation of the complete-polarization sea clutter is realized. Simulation results indicate that complete-polarization sea clutter data obtained through simulation by utilizing the method provided by the invention can reflect sea phenomena such as sea peak, the constraint relations among the polarizationchannels satisfy the physical characteristics of the sea clutter under low sea conditions, and the simulation results can be used for subsequent applications.

The invention discloses a video behavior recognition method and system based on a space enhancement module, and the method comprises the following steps: decoding a to-be-detected video into a frame sequence, and storing the decoded frame sequence in an image form; dividing a video into a plurality of video clips by adopting a sparse sampling strategy, extracting a frame from each video clip, and combining to form a stacked frame sequence; calculating a mean value of three channels of all training video frames in the behavior recognition data set, and subtracting the calculated mean value from the sampled frame image; using a residual neural network 3D-ResNet-18 as a backbone network, and enabling a spatial enhancement module to construct a behavior recognition classification network; setting training parameters, inputting the training set into a behavior recognition classification network for training, and storing trained network parameters; in the model deployment stage, fusing linear operators of the space enhancement module; and inputting a to-be-detected video into the behavior recognition classification network, and outputting a final classification result. According to the invention, the behavior recognition effect is improved, and both effectiveness and universality are achieved.

The method includes steps of pretreatment of iris image, feature description of iris zerocross point, compress and decompress of iris features, and feature matching. The pretreatment includes operations of iris positioning, normalization, image enhancement and noise removing. The pretreatment picks up iris area to be analyzed from original image effectively. Feature of iris zerocross point is analyzed by using multiresolution wavelet so as to build eigenvectors capable of decrypting iris image accurately. Feature compress reduces dimensions of eigenvectors of iris greatly. The compressed eigenvectors needs only small storage space to store. Simple Euclidean distance is adopted in the invention to measure similarity degree between two eigenvectors. Thus, calculated amount is small. The invention possessing fast recognition speed and high recognition accuracy is applicable to application system such as identification and safety guard.

The invention discloses a two-dimensional code image edge detection method and a device. the method comprises steps: binary processing is carried out on an image; feature points on the image are acquired; and the feature points are fit to calculate an edge line. Thus, when the two-dimensional code image edge line is looked up, the operation is simple, complex calculation is not needed, the precision is enhanced through a mathematical model, and the application range is wide.

The invention discloses a direct current side voltage control method of a three-phase four-leg active power filter. The method comprises the following steps: establishing a state space model of the three-phase four-leg active power filter under dq0 coordinates, using a d-axis control quantity to control direct current side voltage, using q-axis and 0-axis control quantities to control alternatingcurrent output current and selecting q-axis and 0-axis output current and the direct current side voltage as system output; defining an error vector, selecting the sliding mode surface and deriving ad-axis current command value; adding the d-axis current command value to output of a harmonic wave detection circuit and taking the sum as total d-axis reference current; and introducing an integral term into the sliding mode surface so as to improve the control precision. When voltage or load of a power grid is unbalanced, two filters are increased, one is a wave trap for filtering out second harmonic waves in voltage at a common connection point, the other one is a low-pass filter for filtering out the harmonic waves contained in a voltage measured value on the direct current side and then accurate compensation of the active power filter can be realized. By adopting the method, the problem of poor effects of conventional PI (proportional and integral) control is solved, and the method further has the advantages of fast response, strong robustness, capability of removing adverse effects of an outer ring on an inner ring and the like.

The invention discloses a segmentation method of a traffic signal lamp. A traffic signal lamp segmentation method based on a one-dimensional Gaussian model is provided to solve a problem of in accurate traffic signal lamp segmentation caused by unreasonable threshold segmentation during traffic signal lamp detection and identification in application of assisted driving and automatic driving and the like. According to the invention, distribution of collected colors of a traffic signal lamp conforms to characteristics of the one-dimensional Gaussian distribution; a segmentation threshold is calculated by using a set confidence; and post processing of a traffic signal lamp segmentation result is carried out by using priori information such as the installation position of the signal lamp and thus a final segmentation result is obtained. With the segmentation method disclosed by the invention, the influence on traffic signal lamp segmentation by the interference factor in a video image canbe suppressed effectively.

The invention relates to an iris identification method based on blanket dimension and lacunarity, belonging to the technical field of iris identification. The method comprises the following main steps of iris identification, iris image normalization, iris feature extraction and pattern matching. In the invention, the blanket dimension and lacunarity of iris texture are respectively calculated, and then two types of feature indexes are combined so as to obtain excellent identification effect; therefore, the method solves the problem that the existing iris identification method is incapable of efficiently utilizing iris fractal geometric feature and radiation physiological property.

The invention belongs to the technical field of radar imaging, and particularly relates to an imaging method for a curve synthetic aperture radar. According to the imaging method, a specific positionof an interpolation region is determined by calculating a geometric center point of a wave number support domain, thus an interpolation area covering a larger wave number support domain range is obtained, data information in the wave number support domain can be well utilized, and the imaging quality is improved. Moreover, the imaging method only relates to simple arithmetic operation, and does not relate to complex calculation operations such as sliding windows, so that the implementation mode of the method is very simple.

An identity recognizing method based on living iris includes such steps as discriminating living iris, quality evaluation and selection of iris image, restoring the out-of-focus image, and iris recognization. Its advantages are high reliability and speed, low requirement of hardware and high correctness.

The present invention discloses a phase modulation code rate estimation method based on peak position, and relates to the code rate technology field. The method concretely comprises the following steps: solving the points where the received phase modulation signal peaks are located, and obtaining a signal peak point sequence; performing linear fitting of the signal peak sequence to obtain a fitting sequence; performing subtraction of the signal peak sequence and the fitting sequence, and obtaining a fitting difference sequence; according to the analysis of the fitting difference sequence, when the phase has abrupt changing, changing the difference sequence value of the corresponding fitting difference sequence, obtaining the distance between the adjacent points of abrupt changing, obtaining the distance minimum value to take as a code element period, and performing reciprocal of the code element period to obtain the code element rate.

The invention belongs to the technical field of radar imaging, and relates to an SAR imaging scene division method based on region-of-interest detection. According to the method, through analyzing the phase error obtained after the echo data of a synthetic aperture radar is preprocessed, and further optimizing a sub-scene division method based on an existing phase gradient sub-scene division method, a region of interest can be well focused, the edge defocusing can be inhibited, and the imaging quality can be improved. In addition, the method does not involve complex calculation operation, and the implementation mode of the method is relatively simple.

The invention provides a spatial point target detection method based on focal plane modulation characteristics. The method comprises the steps that (1) a window scans an infrared image, and direction number indexes, approximating circles, of the image in the window are obtained; (2) according to the direction number, whether the image contained in the window is a background or a target is preliminarily judged; (3) if the image contained in the window is the target, the circle index in the target shape in the window is calculated; (4) a detection threshold is calculated through a circle index; (5) the circle index is compared with the detection threshold, and whether the image contained in the window is the background or the target is judged. According to the spatial point target detection method based on the focal plane modulation characteristics, the modulation characteristics of a focal plane to space weak and small target infrared imaging approximating circles are utilized, target detection and simulation indication are conducted through target shape judgment, noise and the influences of strong points and blind points in the sensor focal plane can be effectively removed, and the method has the good performance for detection of space weak and small targets.

The invention provides a complete-polarization sea clutter simulation method. The complete-polarization sea clutter simulation method comprises the following steps: simulating a random sequence with K distribution by an SIRP (Spherical Invariable Random Process) method and utilizing the random sequence with K distribution as the sea clutter of an HH (Horizontal-Horizontal) polarization channel; then simulating the sea clutter of a VV (Vertical-Vertical) polarization channel by utilizing the amplitude constraint and phase constraint relation between the HH polarization channel and the VV polarization channel; and finally simulating the sea clutter of a cross polarization channel by utilizing the constraint relation between the HH polarization channel and the cross polarization channel, so as to obtain the complete polarization scattering matrix of each scattering unit, so that the simulation of the complete-polarization sea clutter is realized. Simulation results indicate that complete-polarization sea clutter data obtained through simulation by utilizing the method provided by the invention can reflect sea phenomena such as sea peak, the constraint relations among the polarization channels satisfy the physical characteristics of the sea clutter under low sea conditions, and the simulation results can be used for subsequent applications.

The invention provides a transistor grid electrode array with different lengths and a preparation method thereof. According to the method, the exposure process is conduced on two set of masks, in the grid electrode array formed by exposing the first set of the masks, the two end of each grid electrode are flush in the vertical length direction and each grid electrode covers the corresponding grid electrode in a target array at least, in other words, the lengths of all the grid electrodes formed by exposure of the first set of the masks are the same, and the difference of the grid electrode widths after exposure and etching caused by the unequal lengths of the grid electrodes is reduced; the grid electrodes with the same length are made into the grid electrodes with the targeted lengths and the shapes by the adoption of the second set of the masks according to the lengths, due to the fact that the length dimensions of the grid electrodes are ordinarily larger than the width dimensions of the grid electrodes, the difference of the grid electrode lengths after exposure and etching does not exist basically. Accordingly, due to the fact that when the second set of the masks imaging is adopted, the widths of the grid electrodes have been defined by the first set of the masks, the precise control of the widths and the lengths of the grid electrodes is achieved, complex computing is not needed, and achievement is easy.

The invention discloses a signal modulation mode identification method based on peak position discrimination; it includes the following steps: performing peak calculation on received signals to obtain a peak position sequence; performing linear fitting on the peak position sequence, and calculating the difference between the two Sequence; linear fitting of the difference sequence to obtain discriminant parameters; according to the comparison of the discriminant parameters and the set threshold to classify the received signal, and finally identify various signal modulation types and output the judgment result. The principle of the invention is simple, the amount of calculation is small, and it is easy to realize.

A redundant control method based on dual position sensors. The redundant judgment module judges the extreme value faults of the output values ​​of the two position sensors of the servo system, and selects the output value of the position sensor without extreme value faults for output. None of the position sensors have an extreme value fault, then conduct the consistency judgment of the two feedbacks, if the two feedbacks are consistent, keep the current output of the position sensor unchanged; The output value of the position sensor with a zero value fault is output. If the two position sensors do not have a zero value fault, then the feedback and command error judgment is performed, and the output value of the position sensor with a small absolute value of the error value between the output value and the command is output. value to output. The invention has the advantages of simple realization, good fault compatibility and high reliability.

The invention relates to an iris image database synthesis method which is based on the massive texture sampling, comprising the following steps: the pretreatment of an iris image is carried out, and an obtained specific target region is taken as an input image of synthesized textures. The method effectively obtains clean iris textures from the real iris image; the iris image synthesis which is based on the massive texture sampling obtains a virtual synthesized iris image from the real iris image by the non-parametric sampling method; a plurality of frames of similar inner iris images are synthesized from the virtual image by simulating the change characteristics of an iris under various external and internal conditions, thereby generating a large-scale iris image database; the synthesizednormalized iris image is converted to the circular iris; and an evaluation system of the synthesized iris image database is established. The method is used for the large-scale test of the iris recognition algorithm for comparing and predicting the performance of the algorithm and can also be used as the virtual image to carry out the attack test on an existing system, thereby adding the anti-counterfeiting function for an application system.

The present invention proposes a fast positioning method for two-dimensional codes. The method comprises the following steps: S1, obtaining a to-be-processed two-dimensional code image; S2, converting the two-dimensional code image obtained in step S1 into a gray scale image; S3, carrying out median filtering; S4, carrying out edge detection; and S5, carrying out parallel coordinate positioning, wherein the step S5 comprises: S51, mapping the edge image obtained in step S4 to parallel coordinates, S52, rotating the edge image obtained in step S4 along the Y-axis direction, and mapping the rotated edge image to the parallel coordinates; and S53, comparing the distance between each intersection in the image obtained in step S51 with the distance between each intersection in the image obtained in step S52, so that coordinates corresponding to the points with the equal distance jointly describe the area where the two-dimensional code is located. According to the fast positioning method for two-dimensional codes provided by the present invention, in the complex background, fast positioning is carried out on the area where the two-dimensional code is located.

The invention relates to an iris identification method based on blanket dimension and lacunarity, belonging to the technical field of iris identification. The method comprises the following main steps of iris identification, iris image normalization, iris feature extraction and pattern matching. In the invention, the blanket dimension and lacunarity of iris texture are respectively calculated, and then two types of feature indexes are combined so as to obtain excellent identification effect; therefore, the method solves the problem that the existing iris identification method is incapable of efficiently utilizing iris fractal geometric feature and radiation physiological property.

The invention discloses a video behavior recognition method and system based on a spatial enhancement module. The method includes the following steps: decoding the video to be tested into a frame sequence, storing the decoded frame sequence in the form of an image; taking sparse sampling Strategy, divide the video into multiple video clips, extract a frame from each video clip, and combine them into a stacked frame sequence; calculate the mean value of the three channels of all training video frames in the behavior recognition data set, and divide the sampled frames Subtract the calculated mean value from the image; use the residual neural network 3D-ResNet-18 as the backbone network, and construct the behavior recognition and classification network with the space enhancement module; set the training parameters, input the training set into the behavior recognition and classification network for training, and save the training Network parameters; in the model deployment stage, the linear operator of the fusion space enhancement module; the video to be tested is input to the behavior recognition classification network, and the final classification result is output. The invention improves the behavior recognition effect, and has both effectiveness and versatility.

The invention discloses an image recognition algorithm capable of carrying out grade classification on shiitake mushrooms, which is used for solving the problems that the existing shiitake mushroom grade recognition occupies a large amount of manpower, cannot be used in a complex scene, is high in hardware cost and is not suitable for large-scale industrial production. The algorithm comprises the following steps: obtaining a front view image and a rear view image of a shiitake mushroom material; analyzing and processing the front-view image and the rear-view image through a grade recognition algorithm to obtain the quality factor of the mushroom material; Firstly, gradients of front-view and rear-view images of each mushroom material are calculated by using a Sobel operator; binarization processing is performed on gradient images of the front-view image and the rear-view image respectively; and finally, the duty ratio of the gradient image is calculated to obtain a quantized value reflecting the texture strength degree of the material image. The algorithm has the advantages of being simple in principle, high in grading precision, high in processing speed, free of complex calculation, easy to implement in an embedded system and good in industrial application value.

The invention discloses an active power equalization method of an AC-DC-AC converter, and relates to the technical field of electric energy quality management. The method comprises the steps: carrying out the rotation coordinate transformation of a load current, a compensation current and a grid-side current, and obtaining active components and reactive components of the load current, the compensation current and the grid-side current; processing active and reactive components of the load current, and then obtaining a compensation current reference value through rotational coordinate inverse transformation; carrying out PI operation on the compensation current reference value and the compensation current to obtain a primary compensation modulation wave; carrying out PI operation on the active and reactive components of the grid-side current, and then obtaining secondary compensation modulation waves through rotation coordinate transformation. According to the method, the modulation waves of the primary compensation and the modulation waves of the secondary compensation are added, the control signal of a converter is obtained through PWM modulation, control is carried out through the modulation waves of the secondary compensation, grid-side current unbalance can be rapidly compensated, the grid-side current unbalance degree is remarkably reduced, and the reliability and stability of a control system are enhanced.

The invention relates to a driver fatigue detection method based on face video analysis, belonging to the field of vehicle active safety and mode recognition. The method comprises the steps of: carrying out face detection on a face image to be detected for primarily positioning local organs, such as eyes, nose and mouth; further obtaining face characteristic points positioned accurately; and finally, on the basis of an accurate positioning result of multiple frames of face characteristic points to be detected, carrying out quantization description on a face motion characteristic, and obtaining the result of fatigue detection on a driver to be detected according to a face motion statistical indicator. The driver fatigue detection method has high reliability, has no special requirements on hardware, does not involve complex operation, and has good instantaneity so as to be applied in real time in practice. The driver fatigue detection method has better robustness on the aspects of eyeglasses wearing, face angles, nonuniform illumination and the like, and can be suitable for various types of vehicles for reminding a driver of being in a fatigue state so that traffic accidents are avoided.