Image recognition-based electronic loudspeaker
11039253 ยท 2021-06-15
Assignee
Inventors
Cpc classification
H04R1/028
ELECTRICITY
H04R2400/11
ELECTRICITY
H04R2430/01
ELECTRICITY
International classification
Abstract
An image recognition-based electric speaker includes a U-shaped iron for increasing magnetic field and improving anti-magnetism, a magnet installed on the U-shaped iron for generating magnetic field, a washer installed on the magnet for increasing the magnetic field and magnetic permeability, a voice coil installed on the washer for electrically conducting power, a voice coil paper tube installed on the voice coil, a damper installed around the voice coil for maintaining the magnetic gap and improving the capability of withstanding power, a box holder mounted around the low-frequency suspension edge, a gasket installed at the lower edge of the box holder for sealing the box holder airtightly, a low-frequency suspension edge disposed around the mid-frequency vibrating plate, a high-frequency anti-dust cap installed on the voice coil paper tube, and a mid-frequency vibrating plate disposed around the anti-dust cap. This invention has the effect of reducing criminal activities.
Claims
1. An image recognition-based electric speaker, comprising a U-shaped iron, a magnet, a washer, a voice coil, voice coil paper tube, a damper, a box holder, a gasket, a low-frequency suspension edge, a high-frequency dust resisting cover and a mid-frequency vibrating plate, characterized in that the U-shaped iron increases the intensity of magnetic field and improves the outer antimagnetic effect, and the magnet is disposed on the U-shaped iron for generating a magnetic field, and the washer is installed on the magnet for increasing the emphasis of magnetic field and magnetic permeability, and the voice coil is disposed on the washer for electrically conduct the power, and the damper is installed around the voice coil for maintaining the magnetic gap and improving the capability of withstanding power, and the voice coil paper tube is installed on the voice coil, and the anti-dust cap is installed on the voice coil paper tube, and the mid-frequency vibrating plate is installed around the anti-dust cap, and the low-frequency suspension edge is disposed around the mid-frequency vibrating plate, and the box holder is installed around the low-frequency suspension edge, and the gasket is installed at a lower edge of the box holder for sealing the box holder air tightly; a voltage conversion device coupled to the vehicle power supply for converting an output voltage of the vehicle power supply to obtain different supply voltages required by the electric speaker; a brightness sensor installed at a police car rooftop and near the spherical photography device for detecting the ambient brightness of the spherical photography device; an auxiliary lighting source installed at the police car rooftop and near the spherical photography device and coupled to the brightness sensor for receiving the ambient brightness and providing an auxiliary illuminating light based on the ambient brightness which is an image data collected by the spherical photography device; a megaphone, coupled to the embedded processing device in the front-end dashboard of the police car through a cable, for receiving a person's voice in the police car, and amplifying and playing the person's voice through the effect of the U-shaped iron, magnet, washer, voice coil, voice coil paper tube, damper, box holder, gasket, low-frequency suspension edge, high-frequency dust resisting cover and mid-frequency vibrating plate; a spherical photography device, installed on the box holder, for photographing a street view of where the police car is situated, to obtain and output a corresponding high-definition image; a signal analysis device, coupled to the spherical photography device, for receiving a high-definition image, and outputting a mean square error as a target mean square error based on that the pixel value of each pixel of the high-definition image is confirmed to be the pixel value of the high-definition image; a noise analysis device, for receiving high-definition image, and performing a noise analysis of the high-definition image to obtain a primary noise signal with the largest noise amplitude and a secondary noise signal with the second largest noise amplitude, and outputting a signal-to-noise ratio of the high-definition image as the target signal-to-noise ratio based on that the primary noise signal, the secondary noise signal and the high-definition image are confirmed; a filter switching device, coupled to the signal analysis device and the noise analysis device, for receiving the target mean square error and the target signal-to-noise ratio, and if the target signal-to-noise ratio is smaller than or equal to predetermined signal-to-noise ratio threshold and the target mean square error is greater than or equal to the predetermined mean square error threshold, then a first switch signal will be outputted; if the target signal-to-noise ratio is smaller than or equal to the predetermined signal-to-noise ratio threshold and the target mean square error is greater than predetermined mean square error threshold, then a second switch signal will be outputted; if the target signal-to-noise ratio is greater than the predetermined signal-to-noise ratio threshold and the target mean square error is greater than or equal to the predetermined mean square error threshold, then a third switch signal will be outputted; and if the target signal-to-noise ratio is greater than the predetermined signal-to-noise ratio threshold and the target mean square error is smaller than the predetermined mean square error threshold, then a fourth switch signal will be outputted; a Kalman filter device, coupled to the filter switching device, for performing a Kalman filtering of the high-definition image to obtain a target filtered image when receiving the fourth switch signal; a self-adjusting wavelet filter device, coupled to the filter switching device, for performing a self-adjusting wavelet filtering of the high-definition image to obtain a wavelet filtered image and transmitting the wavelet filtered image to the self-adjusting median filtering device when receiving the first switch signal, and performing a self-adjusting wavelet filtering of the high-definition image to obtain the target filtered image directly when receiving the third switch signal; a self-adjusting median filtering device, coupled to the filter switching device, for receiving the wavelet filtered image from the self-adjusting wavelet filter device and performing a self-adjusting median filtering of the wavelet filtered image to obtain the target filtered image when receiving the first switch signal; and when receiving the second switch signal performing a self-adjusting median filtering of the high-definition image to obtain a target filtered image directly; a target recognition device, coupled to the Kalman filter device, the self-adjusting wavelet filter device and the self-adjusting median filtering device, for receiving the target filtered image and matching the target filtered image with various benchmark dangerous conduct appearances or profiles one by one, and if there is a match, then the dangerous conduct signal will be outputted, and if there is no match at all, then a non-dangerous conduct signal will be outputted; a power conversion device, coupled to the voltage conversion device, for confirming the collaborative playback power of the U-shaped iron, magnet, washer, voice coil, voice coil paper tube, damper, box holder, gasket, low-frequency suspension edge, high-frequency dust resisting cover and mid-frequency vibrating plate; an embedded processing device, coupled to the target recognition device, for transmitting a power increasing signal to the power conversion device when receiving the dangerous conduct signal and transmitting a power decreasing signal to the power conversion device when receiving the non-dangerous conduct signal; wherein, the self-adjusting median filtering carried out by the self-adjusting median filtering device comprises: obtaining different types of blocks for each pixel of the received image by using different filtering windows for the pixels and the pixel as the center, confirming the grey variance of the blocks of each type, selecting the filtering window corresponding to the smallest grey variance as a target filtering window, performing a median filtering of the pixel value of the pixel to obtain a filtered pixel value, obtaining the filtered image outputted from the self-adjusting median filtering device according to the filtered pixel values of all pixels of the image; wherein the self-adjusting wavelet filtering performed by the self-adjusting wavelet filter device comprises: performing a wavelet decomposition of the received image to obtain four sub-bands LL, LH, HL, HH, confirming the mean of the four sub-bands, calculating an optical threshold of a wavelet contraction based on the mean, performing a wavelet reconstruction of the image based on the optimal threshold of the wavelet contraction to obtain the filtered image outputted by the self-adjusting wavelet filter device; wherein the Kalman filter device enters from the power saving mode into the operating mode when receiving the fourth switch signal, and the self-adjusting wavelet filter device enters from the power saving mode into the operating mode after receiving the first switch signal or third switch signal, and the self-adjusting median filtering device enters from the power saving mode into the operating mode when receiving the first switch signal or second switch signal.
2. The image recognition-based electric speaker according to claim 1, wherein the self-adjusting wavelet filter device enters from an operating mode into a power saving mode when receiving the second switch signal or the fourth switch signal.
3. The image recognition-based electric speaker according to claim 2, wherein the self-adjusting median filtering device enters from the operating mode into the power saving mode when receiving the third switch signal or the fourth switch signal.
4. The image recognition-based electric speaker according to claim 3, wherein the Kalman filter device enters from the operating mode into the power saving mode when receiving the first switch signal, second switch signal or third switch signal.
Description
BRIEF DESCRIPTION OF THE DRAWING
(1)
(2)
DESCRIPTION OF THE PREFERRED EMBODIMENTS
(3) The technical contents of the present invention will become apparent with the detailed description of preferred embodiments accompanied with the illustration of related drawings as follows. It is intended that the embodiments and figures disclosed herein are to be considered illustrative rather than restrictive.
(4) Common speakers include the following types: Paper cone has the features of natural tone, low price, good rigidity, lightweight material, and high sensitivity and the disadvantages of poor moisture resistance and high difficulty of controlling the consistence of the manufacture. However, the paper cone is often used in high-end HiFi systems because of its good sound output and reversibility. Bulletproof fabric cone has the features of wide frequency response and low distortion, and thus it is the first choice for strong bass lovers, and its disadvantages include high cost, complicated manufacturing process, insufficient sensitivity, and poor effect of light music. Wool knit cone has the features of soft texture, and thus provides excellent performance for soft music and light music and the disadvantages of poor bass effect and lack of strength and shocking power. Polypropylene (PP) cone is popular in high-end speakers and has the features of good consistency, low distortion, and remarkable performance in all aspects. In addition, there are fiber diaphragms and composite diaphragms, and they are expensive and thus are seldom used in common speakers.
(5) However, the present existing police speakers have a single output power control mode which cannot be self-adjusted according to the ambient conditions. To overcome the deficiency of the conventional police speakers, the present invention provides an image recognition-based electric speaker to overcome the aforementioned technical problem of the prior art.
(6) With reference to
(7) The U-shaped iron increases the intensity of magnetic field and improves the outer antimagnetic effect, and the magnet is disposed on the U-shaped iron for generating a magnetic field, and the washer is installed on the magnet for increasing the emphasis of magnetic field and magnetic permeability, and the voice coil is disposed on the washer for electrically conduct the power, and the damper is installed around the voice coil for maintaining the magnetic gap and improving the capability of withstanding power.
(8) The voice coil paper tube is installed on the voice coil, and the anti-dust cap is installed on the voice coil paper tube, and the mid-frequency vibrating plate is installed around the anti-dust cap, and the low-frequency suspension edge is disposed around the mid-frequency vibrating plate, and the box holder is installed around the low-frequency suspension edge, and the gasket is installed at a lower edge of the box holder for sealing the box holder airtightly.
(9) The specific structure of the image recognition-based electric speaker of the present invention will be described in further details below.
(10) The speaker further comprises a voltage conversion device coupled to the vehicle power supply for converting an output voltage of the vehicle power supply to obtain different supply voltages required by the electric speaker.
(11) The speaker further comprises a brightness sensor installed at a police car rooftop and near the spherical photography device for detecting the ambient brightness of the spherical photography device.
(12) The speaker further comprises an auxiliary lighting source installed at the police car rooftop and near the spherical photography device and coupled to the brightness sensor for receiving the ambient brightness and providing an auxiliary illuminating light based on the ambient brightness which is an image data collected by the spherical photography device.
(13) The speaker further comprises:
(14) a megaphone, coupled to the embedded processing device in the front-end dashboard of the police car through a cable, for receiving a person's voice in the police car, and amplifying and playing the person's voice through the effect of the U-shaped iron, magnet, washer, voice coil, voice coil paper tube, damper, box holder, gasket, low-frequency suspension edge, high-frequency anti-dust cap and mid-frequency vibrating plate; a spherical photography device, installed on the box holder, for photographing a street view of where the police car is situated, to obtain and output a corresponding high-definition image; a signal analysis device, coupled to the spherical photography device, for receiving a high-definition image, and outputting a mean square error as a target mean square error based on that the pixel value of each pixel of the high-definition image is confirmed to be the pixel value of the high-definition image;
(15) a noise analysis device, for receiving high-definition image, and performing a noise analysis of the high-definition image to obtain a primary noise signal with the largest noise amplitude and a secondary noise signal with the second largest noise amplitude, and outputting a signal-to-noise ratio of the high-definition image as the target signal-to-noise ratio based on that the primary noise signal, the secondary noise signal and the high-definition image are confirmed;
(16) a filter switching device, coupled to the signal analysis device and the noise analysis device, for receiving the target mean square error and the target signal-to-noise ratio, and if the target signal-to-noise ratio is smaller than or equal to predetermined signal-to-noise ratio threshold and the target mean square error is greater than or equal to the predetermined mean square error threshold, then a first switch signal will be outputted; if the target signal-to-noise ratio is smaller than or equal to the predetermined signal-to-noise ratio threshold and the target mean square error is greater than predetermined mean square error threshold, then a second switch signal will be outputted; if the target signal-to-noise ratio is greater than the predetermined signal-to-noise ratio threshold and the target mean square error is greater than or equal to the predetermined mean square error threshold, then a third switch signal will be outputted; and if the target signal-to-noise ratio is greater than the predetermined signal-to-noise ratio threshold and the target mean square error is smaller than the predetermined mean square error threshold, then a fourth switch signal will be outputted;
(17) a Kalman filter device, coupled to the filter switching device, for performing a Kalman filtering of the high-definition image to obtain a target filtered image when receiving the fourth switch signal;
(18) a self-adjusting wavelet filter device, coupled to the filter switching device, for performing a self-adjusting wavelet filtering of the high-definition image to obtain a wavelet filtered image and transmitting the wavelet filtered image to the self-adjusting median filtering device when receiving the first switch signal, and performing a self-adjusting wavelet filtering of the high-definition image to obtain the target filtered image directly when receiving the third switch signal;
(19) a self-adjusting median filtering device, coupled to the filter switching device, for receiving the wavelet filtered image from the self-adjusting wavelet filter device and performing a self-adjusting median filtering of the wavelet filtered image to obtain the target filtered image when receiving the first switch signal; and when receiving the second switch signal performing a self-adjusting median filtering of the high-definition image to obtain a target filtered image directly;
(20) a target recognition device, coupled to the Kalman filter device, the self-adjusting wavelet filter device and the self-adjusting median filtering device, for receiving the target filtered image and matching the target filtered image with various benchmark dangerous conduct appearances or profiles one by one, and if there is a match, then the dangerous conduct signal will be outputted, and if there is no match at all, then a non-dangerous conduct signal will be outputted;
(21) a power conversion device, coupled to the voltage conversion device, for confirming the collaborative playback power of the U-shaped iron, magnet, washer, voice coil, voice coil paper tube, damper, box holder, gasket, low-frequency suspension edge, high-frequency anti-dust cap and mid-frequency vibrating plate;
(22) an embedded processing device, coupled to the target recognition device, for transmitting a power increasing signal to the power conversion device when receiving the dangerous conduct signal and transmitting a power decreasing signal to the power conversion device when receiving the non-dangerous conduct signal;
(23) wherein, the self-adjusting median filtering carried out by the self-adjusting median filtering device comprises: obtaining different types of blocks for each pixel of the received image by using different filtering windows for the pixels and the pixel as the center, confirming the grey variance of the blocks of each type, selecting the filtering window corresponding to the smallest grey variance as a target filtering window, performing a median filtering of the pixel value of the pixel to obtain a filtered pixel value, obtaining the filtered image outputted from the self-adjusting median filtering device according to the filtered pixel values of all pixels of the image; wherein the self-adjusting wavelet filtering performed by the self-adjusting wavelet filter device comprises: performing a wavelet decomposition of the received image to obtain four sub-bands LL, LH, HL, HH, confirming the mean of the four sub-bands, calculating an optical threshold of a wavelet contraction based on the mean, performing a wavelet reconstruction of the image based on the optimal threshold of the wavelet contraction to obtain the filtered image outputted by the self-adjusting wavelet filter device;
(24) wherein the Kalman filter device enters from the power saving mode into the operating mode when receiving the fourth switch signal, and the self-adjusting wavelet filter device enters from the power saving mode into the operating mode after receiving the first switch signal or third switch signal, and the self-adjusting median filtering device enters from the power saving mode into the operating mode when receiving the first switch signal or second switch signal.
(25) In the speaker, the self-adjusting wavelet filter device enters from an operating mode into a power saving mode when receiving the second switch signal or the fourth switch signal.
(26) In the speaker, the self-adjusting median filtering device enters from the operating mode into the power saving mode when receiving the third switch signal or the fourth switch signal.
(27) In the speaker, the Kalman filter device enters from the operating mode into the power saving mode when receiving the first switch signal, second switch signal or third switch signal.
(28) In addition, the image filtering suppresses the noise as shown in the target icon in the figure while maintaining the detailed characteristics of the image as much as possible, and this is a necessary operation in an image pre-processing process, and the effect of the processing directly affects the validity and reliability of the subsequent image processing and analysis.
(29) Due to the imperfections in imaging systems, transmission media and recording devices, the formation and transmitting process of digital images are often affected by various types of noises. In addition, noises may be introduced into the resulted image in some imaging processing cases, if the inputted image object is not as expected. These noises are often expressed in form of an isolated pixel or block of an image having a strong visual effect. In general, a noise signal appears as useless information with respect to a studied object but it will disturb the observable information of the image. Digital image signals and noises are in the maximum or minimum values, and these extremum values may cause bright or dark spots of an image and lower the image quality significantly through the addition or subtraction of these extremum values on the real grey value of the image pixel, or even affects the restoration, division, characteristic fetching, image identification of the image. It is necessary to take the following two basic factors into consideration on the effect of suppressing noises effectively: The noises in the target and background must be removed effectively. In addition, the target shape, size, and specific geometrically and topologically structural characteristics of the image must be protected properly.
(30) One of the common image filtering modes is a nonlinear filter, generally. If a signal spectrum is mixed and overlapped with a noise spectrum or a signal contains a non-superimposing noise, such as the existence of a noise caused by system linearity or a non-Gaussian noise, the traditional linear filtering technology such as Fourier transform will express the image in certain fuzzy image details (such as an edge) while filtering the noise. As a result, the positioning precision and extractability of the linear characteristic of the image will be reduced. The nonlinear filter is a nonlinear mapping of the input signal and often maps a certain specific noise to zero while maintaining the desired characteristics of the signal, and thus it can overcome the deficiencies of the linear filter to a certain extent.
(31) Compared with the traditional speakers with a fixed output power, the image recognition-based electric speaker of the present invention integrates a plurality of high-precision image processing devices into the traditional speakers to confirm whether or not any nearby dangerous conduct exists and to perform a self-adjustment of the output power of the speaker, so as to enhance the automation level of the speaker.
(32) While the present invention has been described by means of specific embodiments, numerous modifications and variations could be made thereto by those skilled in the art without departing from the scope and spirit of the present invention set forth in the claims.