32results about How to "Suppress ambient noise" patented technology

The invention provides a system and a method for reducing environmental noise as well as a device applying the system, and is particularly applied to a stereo radio module with at least two input terminals. The method comprises the following steps: respectively receiving audio frequency including a main audio frequency part and most part of the environmental noise; executing signal calibration in the system so as to reduce the difference between the input terminals; respectively obtaining an environmental noise part with less main audio frequency and a main audio frequency part with less environmental noise by virtue of adaptive beam forming and speech extraction technologies; after frequency domain transformation, estimating the environmental noise received by the system by virtue of a nonlinear noise suppression technology; obtaining a gain; denoising by means of the gain; and finally generating the continuously output audio frequency through carrying out domain transformation again. The system and the method as well as the device applying the same have the beneficial effects of effectively suppressing the environmental noise during call, improving the intelligibility and comfortability during the audio or speech communication process, preventing the call from being affected by too much noise, ensuring instant signal calibration, and maximally reducing the noise included in the audio frequency.

The invention discloses a wave arrival estimation method based on the high-order accumulated amount, and belongs to the technical field of signal processing. According to the method, observation signals received by an equal-interval linear sensor array are used to estimate the direction of arrival and arrival time of a signal source. The method comprises the following steps that 1) spatial-time domain smoothing is carried out on the observation signals after Fourier Transform; 2) a four-order accumulated amount matrix of the signals after spatial-time domain smoothing is constructed; 3) an iterative part SVD method is used to construct a signal subspace and a noise subspace of the observation signals according to the four-order accumulated amount matrix; and 4) according to the orthogonality between the signal subspace and the noise subspace of the observation signals, the direction of arrival and arrival time of the signal source are estimated. The invention also discloses an ocean acoustic tomography method based on sound ray spreading time tomography and a positioning method. According to the invention, the computing complexity of an algorithm can be greatly reduced, the instantaneity of the algorithm is improved, and consumption of hardware resource is reduced.

An underwater acoustic channel tracking and prediction method based on inertial information, which relates to the field of signal processing technology. In view of the prior art, which is restricted by the marine environment and the working conditions of underwater vehicles, the propagation of sound waves in water will be affected by high noise and complex Due to the influence of multi-channel expansion and Doppler frequency shift, it is difficult to maintain uninterrupted and reliable communication, which affects the performance of underwater acoustic positioning. The ability to maintain the short-term accuracy of the guide can predict the current acoustic channel and accumulate multi-channel energy, which will help the underwater acoustic receiver to suppress environmental noise, improve the accuracy of parameter estimation, maintain the connectivity of the acoustic link, and improve the robustness of ultra-short baseline positioning sex.

The invention discloses a target recognition method and device. The object recognition method includes the following steps: using a linear feature detection method to perform linear feature detection on the image to be recognized to obtain the detected linear feature, and the linear feature detection method converts the detection of the linear feature of the image space into the detection of another spatial extreme point And wherein a continuous clustering operator is used to make the continuous image points have a greater weight than the discrete image points in the conversion; and according to the characteristics of the target to be recognized in the image to be recognized, the detected linear features are identified Target. Using the target recognition method and device according to the present invention, when detecting and identifying weak linear targets from high-resolution remote sensing images, system noise and environmental noise can be effectively suppressed, targets can be successfully identified and extracted, misjudgments can be prevented, and detection short line segment.

The invention relates to an onsite classroom pickup and loudspeaking system adapted to the position of a speaker. The system comprises a pickup unit, a loudspeaker array and a processing circuit, wherein the pickup unit converts obtained audio signals of the speaker into electrical signal based on the position of the speaker, the processing circuit removes noises of the electrical signals and amplifies the de-noised signals to generate driving signals, and the loudspeaker array plays sounds of the speaker according to the driving signals. The system of the invention needs only one pickup unit, the speaker can walk in the classroom to teach class needless of wearing extra devices, local environmental noises can be effectively inhibited, and the problem of howling is eliminated.

The invention discloses a voice enhancement method, system and communication equipment. Wherein, the method includes: acquiring a voice input signal; performing time-frequency compensation on the voice input signal; calculating a plurality of voice feature information after time-frequency compensation, and classifying the voice input signal to obtain a voice classification result; according to the multiple voice feature information and the speech classification result to determine a speech masking value; based on the speech masking value, determine a speech output signal. The invention solves the technical problem that the voice enhancement method in the related art adopts a fixed mask value, which will leave a lot of noise or damage the voice of the target object, resulting in poor call effect.

The invention relates to the field of speech enhancement, and discloses a quaternary differential microphone array super directivity frequency invariant beamforming method. An output signal of a microphone array is decomposed into a plurality of sub-frequency bands according to a frequency band after the output signal is collected, after compensation or inhibition, reconstitution and synthesis areconducted to obtain a target signal, and then directivity adjustment is conducted on the microphone array, a maximum directivity factor is obtained, maximum white noise suppression capability of themicrophone array is obtained, then the target signal is output after directivity treatment and white noise suppression treatment are conducted through the microphone array, then a matrix vector is designed, and the output signal is subjected to speech signal enhancement and noise signal suppression. According to the quaternary differential microphone array super directivity frequency invariant beamforming method.

The invention relates to a method to operate a device in a sterile environment and a device suitable for a sterile environment. The device is controlled by a receiving device for acquiring contact-free user input, wherein the user input is executed by different users (X, X+1). The method includes the following steps: a) connecting a first operation mode (Z1) after any of the contact-free user input of a first user or a second user is acquired; b) switching from the first operation mode to a second operation mode (Z2) after the preset contact-free user input is acquired, wherein the preset contact-free user input during the switching from the first operation mode to the second operation mode is only the contact-free user input executed in the first operation mode. Other operation modes (Z0) are activated after a non-activated stage of the user input is identified and / or other preset contact-free user input of the user of the preset user input executed in b) is acquired.