Object detection device and method of operation thereof

Abstract

An object detection device can include a transducer configured to convert a transmission signal emitted toward an object into a digital transmission signal and a reception signal reflected from the object into a digital reception signal according to a predetermined sampling period, and at least one processor configured to interpolate between elements of the digital transmission signal and the digital reception signal having the predetermined sampling period to obtain an interpolated transmission signal and an interpolated reception signal, remove noise from each of the interpolated transmission signal and the interpolated reception signal, generate a cross-correlation signal between the interpolated transmission signal from which the noise is removed and the interpolated reception signal from which the noise is removed, and obtain a three-dimensional (3D) image of the object based on at least one peak of the cross-correlation signal.

Description

[0001] Cross-references to related applications

[0002] This application claims the benefit of U.S. Provisional Application No. 62 / 953,755, filed December 26, 2019, with the U.S. Patent and Trademark Office, and Korean Patent Application No. 10-2020-00040467, filed April 2, 2020, with the Korean Intellectual Property Office, the disclosures of which are incorporated herein by reference in their entirety. Technical Field

[0003] The apparatus and methods consistent with the example embodiments relate to object detection based on optical signals. Background Technology

[0004] Object detection devices can generate three-dimensional (3D) images of objects by measuring the time-of-flight (ToF) of light relative to the object. Specifically, the object detection device can calculate the distance to the object by measuring the return time until the light signal emitted from the light source is reflected by the object, and can generate a depth image of the object based on the calculated distance.

[0005] Object detection devices calculate Time-of-Flight (ToF) by converting optical signals into digital signals; however, existing object detection devices only use analog-to-digital converters (ADCs) to convert optical signals into digital signals. Therefore, the achievable distance resolution is limited. Furthermore, existing object detection devices may not provide solutions for removing low-frequency noise. Summary of the Invention

[0006] Example embodiments provide an object detection device and its operating method that reduce the possibility of false detection while increasing distance resolution.

[0007] The technical problems to be solved are not limited to those described above, and other technical problems can be inferred from the following embodiments.

[0008] According to one aspect of an example embodiment, an object detection apparatus is provided, comprising: a converter configured to convert a transmitted signal emitted toward an object into a digital transmitted signal and a received signal reflected from the object into a digital received signal according to a predetermined sampling period; and at least one processor configured to: interpolate between elements of the digital transmitted signal and the digital received signal having a predetermined sampling period to obtain an interpolated transmitted signal and an interpolated received signal; remove noise from each of the interpolated transmitted signal and the interpolated received signal; generate a cross-correlation signal between the noise-removed interpolated transmitted signal and the noise-removed interpolated received signal; and acquire a three-dimensional (3D) image of the object based on at least one peak of the cross-correlation signal.

[0009] The converter can be configured to output each of the transmission signal and the reception signal by converting each of the transmission signal and the reception signal into vector data in the form of a column vector or a row vector.

[0010] The at least one processor can be further configured to interpolate between each element included in the vector data.

[0011] The at least one processor can be further configured to remove high-frequency noise by accumulating each element included in the interpolated vector data over a predetermined time and outputting an average value of the accumulated elements, wherein the high-frequency noise can be a part of noise having a frequency higher than a predetermined upper threshold value.

[0012] The at least one processor can be further configured to receive the vector data from which the high-frequency noise is removed as first input data, generate second input data by moving each first element included in the first input data by a predetermined size in a predetermined direction, and output third input data from which low-frequency noise is removed by subtracting each second element included in the second input data from each first element included in the first input data, wherein the low-frequency noise can be a part of noise having a frequency lower than a predetermined lower threshold value.

[0013] The at least one processor can be further configured to insert zero in the third input data when there is no second element corresponding to any first element of the first input data in the second input data.

[0014] The at least one processor can be further configured to detect at least one peak value from the cross-correlation signal, determine a quality of the cross-correlation signal based on the at least one peak value, and acquire a 3D image of the object based on the quality of the cross-correlation signal.

[0015] The at least one processor can be further configured to detect a first peak value having a greatest absolute value among the at least one peak value of the cross-correlation signal, detect a second peak value having a second greatest absolute value among the at least one peak value of the cross-correlation signal, and determine the quality of the cross-correlation signal based on the absolute value of the first peak value and the absolute value of the second peak value.

[0016] The at least one processor can be further configured to generate a point cloud based on the cross-correlation signal having a quality greater than or equal to a predetermined reference quality, and acquire a 3D image of the object based on the generated point cloud.

[0017] According to an aspect of another example embodiment, there is provided a method for object detection, the method including: converting a transmission signal transmitted toward an object and a reception signal reflected from the object into digital transmission and reception signals according to a predetermined sampling period; interpolating between elements of the digital transmission and reception signals having the predetermined sampling period to obtain interpolated transmission and reception signals; removing noise from each of the interpolated transmission and reception signals; generating a cross-correlation signal between the interpolated transmission signal from which the noise is removed and the interpolated reception signal from which the noise is removed; and obtaining a three-dimensional (3D) image of the object based on at least one peak value of the cross-correlation signal.

[0018] The converting can include converting each of the transmission and reception signals into vector data in a column vector or a row vector form, and the interpolating can include interpolating between each element included in the vector data.

[0019] The removing noise can include removing high frequency noise by accumulating each element included in the interpolated vector data over a predetermined time and outputting an average value of the accumulated elements, receiving the vector data from which the high frequency noise is removed as first input data, generating second input data by moving each first element included in the first input data by a predetermined size in a predetermined direction, and outputting third input data from which low frequency noise is removed by subtracting each second element included in the second input data from each first element included in the first input data, wherein the high frequency noise is a part of noise having a frequency higher than a predetermined upper threshold value.

[0020] The outputting the third input data can include inserting zero in the third input data when there is no second element corresponding to any first element of the first input data in the second input data.

[0021] The obtaining can include detecting at least one peak value in the cross-correlation signal, determining a quality of the cross-correlation signal based on the at least one peak value, and obtaining the 3D image of the object based on the quality of the cross-correlation signal.

[0022] The detecting can include detecting a first peak value having a largest absolute value among the at least one peak value of the cross-correlation signal, and detecting a second peak value having a second largest absolute value among the at least one peak value of the cross-correlation signal. The determining can include determining the quality of the cross-correlation signal based on the absolute value of the first peak value and the absolute value of the second peak value. The obtaining can include generating a point cloud based on the cross-correlation signal whose quality is greater than or equal to a predetermined reference quality, and obtaining the 3D image of the object based on the generated point cloud.

[0023] According to an aspect of another example embodiment, there is provided an object detection device including: a converter configured to convert a plurality of analog transmission signals to be transmitted to an object into a plurality of digital transmission signals according to a first sampling period, and to convert a plurality of analog reception signals reflected from the object into a plurality of digital reception signals according to the first sampling period; and at least one processor configured to: interpolate the plurality of digital transmission signals and the plurality of digital reception signals according to a second sampling period; combine the interpolated plurality of digital transmission signals with each other to obtain a first combined signal and combine the interpolated plurality of digital reception signals with each other to obtain a second combined signal; generate a cross-correlation signal based on the first combined signal and the second combined signal; and acquire a three-dimensional (3D) image of the object based on the cross-correlation signal.

[0024] The at least one processor can be further configured to remove noise having a frequency greater than a predetermined upper threshold from the interpolated plurality of digital transmission signals and the interpolated plurality of digital reception signals by combining the interpolated plurality of digital transmission signals with each other and the interpolated plurality of digital reception signals with each other, respectively.

[0025] The at least one processor can be further configured to move the first combined signal by a first predetermined size in a first predetermined direction, move the second combined signal by a second predetermined size in a second predetermined direction, and generate the cross-correlation signal based on the moved first combined signal and the moved second combined signal.

[0026] The at least one processor can be further configured to remove noise having a frequency less than a predetermined lower threshold from the first combined signal and the second combined signal by moving the first combined signal and the second combined signal, respectively. BRIEF DESCRIPTION OF DRAWINGS

[0027] The above and / or other aspects will be more apparent by describing certain example embodiments, with reference to the accompanying drawings, in which:

[0028] Figure 1 is a diagram for explaining an operation of an object detection device according to an example embodiment;

[0029] Figure 2 is an internal block diagram of an object detection device according to an example embodiment;

[0030] Figure 3 is a diagram for explaining a method of interpolating transmission signals and reception signals according to an example embodiment;

[0031] Figure 4 is a diagram for explaining an effect of increasing a distance resolution according to interpolation of transmission signals and reception signals according to an example embodiment;

[0032] Figure 5is a graph for explaining an effect of increasing distance resolution according to interpolation of a transmission signal and a reception signal according to an example embodiment;

[0033] Figure 6 is a graph for explaining a method of removing low-frequency noise of a transmission signal and a reception signal according to an example embodiment;

[0034] Figure 7 is a graph for explaining an effect generated when zero is inserted in a blank position of third input data according to an example embodiment;

[0035] Figure 8A and Figure 8B is a graph for explaining an effect of increasing a detection distance according to removal of low-frequency noise of a transmission signal and a reception signal according to an example embodiment;

[0036] Figure 9 is a graph for explaining an effect of increasing a detection distance according to removal of low-frequency noise of a transmission signal and a reception signal according to an example embodiment;

[0037] Figure 10 is a graph for explaining a method of calculating a quality of a cross-correlation signal according to an example embodiment;

[0038] Figure 11 is a graph for explaining an effect of preventing false detection according to quality calculation of a cross-correlation signal according to an example embodiment;

[0039] Figure 12 is a flowchart of an object detection method according to an example embodiment;

[0040] Figure 13 is a flowchart of a method of removing high-frequency noise in a method of Figure 12 according to an example embodiment;

[0041] Figure 14 is a flowchart of a method of removing low-frequency noise in a method of Figure 12 according to an example embodiment; and

[0042] Figure 15 is a flowchart of a method of acquiring a 3D image in a method of Figure 12 according to an example embodiment. DETAILED DESCRIPTION

[0043] Example embodiments are described in more detail below with reference to accompanying drawings.

[0044] In the following description, like reference numerals are used to refer to like elements throughout the several views. Matters defined in the description such as detailed construction and elements are to help a full understanding of example embodiments. However, it is apparent that example embodiments can be practiced without those specific defined matters. Also, since well-known functions or constructions will make the description unclear, they are not described in detail.

[0045] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items. Expressions such as "at least one of," when preceding the list of elements, modify the entire list of elements and do not modify the individual elements of the list. For example, the expression "at least one of a, b, and c" should be understood as including only a, only b, only c, both a and b, both a and c, both b and c, all of a, b, and c, or any variations of the above examples.

[0046] The phrase "in some embodiments" or "in an embodiment" appearing in various places in this specification does not necessarily refer to the same embodiment.

[0047] Some embodiments of the present disclosure can be represented as functional block configurations and various processing steps. Some or all of the functional blocks can be implemented in various numbers of hardware and / or software configurations that perform specific functions. For example, the functional blocks of the present disclosure can be implemented by one or more microprocessors or by circuit configurations for a given function. Also, for example, the functional blocks of the present disclosure can be implemented in various programming or scripting languages. The functional blocks can be implemented in algorithms that run on one or more processors. Also, the present disclosure can employ existing technologies for electronic configurations, signal processing, and / or data processing. Terms such as "mechanism," "element," "device," and "configuration" can be used broadly and are not limited to mechanical and physical configurations.

[0048] Also, connection lines or connection members between components shown in the drawings are merely examples of functional and / or physical or circuit connections. In actual devices, connections between components can be represented by various functional connections, physical connections, or circuit connections that can be replaced or added.

[0049] Figure 1 FIG. 1 is a diagram for explaining an operation of an object detection device 100 according to an example embodiment.

[0050] Referring to Figure 1The object detection device 100 includes a transmitter 110 configured to emit a transmission signal S1 toward an object OBJ, a receiver 120 configured to receive a reception signal S2 reflected from the object OBJ, and a controller 130 configured to control the transmitter 110 and the receiver 120. The controller 130 can include at least one processor and at least one memory.

[0051] The object detection device 100 can be a three-dimensional (3D) sensor that generates a 3D image of the object OBJ. For example, the object detection device 100 can include a light detection and ranging (LiDAR), a radar, or the like, but is not limited thereto.

[0052] The transmitter 110 can output light to be used for analyzing a position, a shape, or the like of the object OBJ. For example, the transmitter 110 can output light having a wavelength of an infrared band. When light in the infrared band is used, mixing with natural light in a visible light band including sunlight can be prevented. However, it is not necessarily limited to the infrared band, and the transmitter 110 can emit light of various wavelength bands.

[0053] The transmitter 110 can include at least one light source. For example, the transmitter 110 can include a light source such as a laser diode (LD), an edge emitting laser, a vertical cavity surface emitting laser (VCSEL), a distributed feedback laser, a light emitting diode (LED), a super luminescent diode (SLD), or the like.

[0054] The transmitter 110 can also generate and output light of a plurality of different wavelength bands. Also, the transmitter 110 can generate and output pulsed light or continuous light. The light generated by the transmitter 110 can be emitted as the transmission signal S1 toward the object OBJ.

[0055] According to an example embodiment, the transmitter 110 can further include a beam control device for changing an emission angle of the transmission signal S1. For example, the beam control device can be a scanning mirror or an optical phased array.

[0056] The controller 130 can control the transmitter 110 to change the emission angle of the transmission signal S1. The controller 130 can control the transmitter 110 so that the transmission signal S1 scans the entire object OBJ. In one embodiment, the controller 130 can control the transmitter 110 so that the transmission signal S1 output from each of the plurality of light sources scans the object OBJ at different emission angles. In another example embodiment, the controller 130 can control the transmitter 110 so that the transmission signal S1 output from each of the plurality of light sources scans the object OBJ at the same emission angle.

[0057] Receiver 120 may include at least one light detection element, and the light detection element may individually detect and receive the received signal S2 reflected from object OBJ. According to an example embodiment, receiver 120 may also include optical elements for collecting the received signal S2 onto the predetermined light detection element.

[0058] The predetermined light detection element can be a sensor capable of sensing light, and can be, for example, a light receiving device configured to generate an electrical signal through light energy. There are no particular limitations on the type of light receiving device.

[0059] The controller 130 can perform signal processing to acquire information about the object OBJ using the received signal S2 detected by the receiver 120. The controller 130 can determine the distance to the object OBJ based on the time-of-flight of the light output by the transmitter 110 and perform data processing to analyze the position and shape of the object OBJ. For example, the controller 130 can generate a point cloud based on the distance information to the object OBJ and acquire a 3D image of the object OBJ based on the point cloud.

[0060] The 3D images acquired by controller 130 can be sent to another unit for its use. For example, the information can be sent to the controller of an autonomous vehicle or drone, in which object detection device 100 is employed. In addition to the above, the information can be used in smartphones, mobile phones, personal digital assistants (PDAs), laptops, personal computers (PCs), wearable devices, and other mobile or non-mobile computing devices.

[0061] In addition, the controller 130 can control the overall operation of the object detection device 100, including the control of the transmitter 110 and the receiver 120. For example, the controller 130 can perform power control, on / off control, pulse wave (PW) or continuous wave (CW) generation control for the transmitter 110.

[0062] Apart from Figure 1 In addition to the components disclosed herein, the object detection device 100 may also include other general components.

[0063] For example, the object detection device 100 may also include a memory for storing various types of data. The memory may store data that is processed by the object detection device 100 and data that will be processed. Furthermore, the memory may store applications, drivers, etc., that will be driven by the object detection device 100.

[0064] The memory can include random access memory (RAM) (e.g., dynamic random access memory (DRAM) and static random access memory (SRAM)), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), a CD-ROM, a Blu-ray or other optical disk storage device, a hard disk drive (HDD), a solid state drive (SSD), or a flash memory, and can also include other external storage devices that can be accessed by the object detection device 100.

[0065] Figure 2 is an internal block diagram of the object detection device 100 according to an example embodiment.

[0066] Referring to Figure 2 , the object detection device 100 can include a transmitter 110, a receiver 120, a converter 140, an interpolation unit 150, a filter unit 160, and a controller 130. The interpolation unit 150, the filter unit 160, and the controller 130 can be implemented by one or more processors.

[0067] The transmitter 110 can output a transmission signal toward an object. Also, the transmitter 110 can output a portion of the transmission signal to the converter 140. The portion of the transmission signal can be used to calculate a time of flight of light. The receiver 120 can receive a reception signal reflected from the object. Figure 2 The transmitter 110 of the object detection device 100 can correspond to Figure 1 The transmitter 110 of the object detection device 100, Figure 2 The receiver 120 of the object detection device 100 can correspond to Figure 1 The receiver 120 of the object detection device 100.

[0068] The converter 140 can convert each of the transmission signal and the reception signal into a digital signal according to a predetermined sampling period. For example, the sampling period can be set in a range of 100 Hz to 10 GHz. The converter 140 can output the digital signal as vector data in a column vector or a row vector form. The vector data can represent an array in a column vector or a row vector form including a set of elements. Quantized values of the transmission signal and the reception signal can be stored in each element.

[0069] Specifically, the converter 140 can include a first analog-to-digital converter 141 and a second analog-to-digital converter 142.

[0070] The first analog-to-digital converter 141 can convert the transmission signal into a first digital signal based on the predetermined sampling period. In other words, the first digital signal can be a converted transmission signal. The first analog-to-digital converter 141 can output the first digital signal in a column vector or a row vector form. Quantized values of the transmission signal according to the sampling period can be stored in each element.

[0071] The second analog-digital converter 142 can convert the reception signal into a second digital signal based on a predetermined sampling period. In other words, the second digital signal can be a converted reception signal. The second analog-digital converter 142 can output the second digital signal in a column vector or a row vector form. Quantized values of the reception signal according to the sampling period can be stored in each element.

[0072] The first analog-digital converter 141 can output the first digital signal to a first resampling unit 151 included in the interpolation unit 150. The second analog-digital converter 142 can output the second digital signal to a second resampling unit 152 included in the interpolation unit 150.

[0073] The interpolation unit 150 can interpolate the first digital signal and the second digital signal by predicting elements between sampling periods. The interpolation unit 150 can interpolate between each element included in the vector data. For example, the interpolation unit 150 can interpolate the first digital signal and the second digital signal by using at least one of a linear interpolation method, a polynomial interpolation method, a spline interpolation method, an exponential interpolation method, a logarithmic linear interpolation method, a Lagrange interpolation method, a Newton interpolation method, and a bilinear interpolation method. However, the above-described interpolation methods are merely examples, and various interpolation methods can be used to interpolate the first digital signal and the second digital signal.

[0074] The interpolation unit 150 can interpolate between sampling periods according to a predetermined interpolation period. The predetermined interpolation period can be set in a range of 2 to 20. In other words, the interpolation unit 150 can interpolate the first digital signal and the second digital signal by dividing the sampling period into 2 to 20 parts. For example, when the sampling period is 100 Hz, the interpolation unit 150 can interpolate the first digital signal and the second digital signal by predicting elements corresponding to 50 Hz.

[0075] The interpolation unit 150 can include the first resampling unit 151 and the second resampling unit 152. The first resampling unit 151 can interpolate the first digital signal, and the second resampling unit 152 can interpolate the second digital signal. In an example embodiment, resampling (e.g., upscaling such as interpolation) can be performed on a sampling signal before integrating, accumulating, or combining the sampling signal to obtain additional information and improve a distance resolution.

[0076] The first resampling unit 151 can output the interpolated first digital signal to the filter unit 160, and the second resampling unit 152 can output the interpolated second digital signal to the filter unit 160.

[0077] The interpolated first digital signal can be referred to as an interpolated transmission signal, and the interpolated second digital signal can be referred to as an interpolated reception signal.

[0078] Since the object detection device 100 of the disclosure increases the sampling rate through the interpolation unit 150, the distance resolution is significantly improved without changing the hardware design (e.g., using an analog-to-digital converter having a high sampling rate).

[0079] The filter unit 160 can remove noise of each of the interpolated transmission signal and the interpolated reception signal. The filter unit 160 can remove high-frequency noise of the interpolated transmission signal and the interpolated reception signal. Also, the filter unit 160 can remove low-frequency noise of the interpolated transmission signal and the interpolated reception signal. In particular, the filter unit 160 can remove one or more portions of the interpolated transmission signal and the interpolated reception signal, in which the frequency is greater than a predetermined upper threshold or less than a predetermined lower threshold, from the interpolated transmission signal and the interpolated reception signal. To this end, the filter unit 160 can include a first denoising unit 161 and a second denoising unit 162. For example, the filter unit 160 can suppress or remove high-frequency noise by integration, and can suppress or remove low-frequency noise by differentiation.

[0080] The first denoising unit 161 can remove high-frequency noise of the interpolated transmission signal and the interpolated reception signal. According to an example embodiment, the first denoising unit 161 can remove only high-frequency noise of the interpolated reception signal.

[0081] In particular, since the converter 140 outputs the transmission signal and the reception signal in the form of vector data, and the interpolation unit 150 interpolates between each element included in the vector data, the first denoising unit 161 can receive the interpolated transmission signal in the form of vector data.

[0082] The first denoising unit 161 can accumulate or combine each element included in the interpolated vector data for a predetermined time. Also, the first denoising unit 161 can remove high-frequency noise by outputting an average value of the accumulated elements. At this time, the predetermined time can be 0.01 ms, but is not limited thereto.

[0083] For example, when the first denoising unit 161 receives first row vector data and second row vector data for a predetermined time, in which the elements included in the first row vector data are 2, 4, 7, and 9, respectively, and the elements included in the second row vector data are 3, 5, 4, and 9, respectively, the first denoising unit 161 can obtain accumulated row vector data in which the elements are 5, 9, 11, and 18. Also, the first denoising unit 161 can output average row vector data in which the average elements of the accumulated row vector data are 2.5, 4.5, 5.5, and 9.

[0084] When the first de-noising unit 161 accumulates the vector data for a predetermined time and outputs an average value of the vector data, high-frequency noise of the interpolated transmission signal and the interpolated reception signal can be removed.

[0085] The second de-noising unit 162 can remove low-frequency noise of the interpolated transmission signal and the interpolated reception signal.

[0086] Specifically, the second de-noising unit 162 can receive the vector data from which the high-frequency noise is removed as first input data. In addition, the second de-noising unit 162 can generate second input data by performing element movement on each element included in the first input data in a predetermined direction and a predetermined size.

[0087] The predetermined direction can be any one of left, right, up, and down directions. When the vector data is a row vector, the predetermined direction can be set to left or right. At this time, the left side can mean a direction in which the column address of the vector data decreases, and the right side can mean a direction in which the column address of the vector data increases. Also, when the vector data is a column vector, the predetermined direction can be set to the upper side or the lower side. At this time, the upper side can mean a direction in which the row address of the vector data decreases, and the lower side can mean a direction in which the row address of the vector data increases. The predetermined size can be set to 500 elements.

[0088] The second de-noising unit 162 can output third input data from which the low-frequency noise is removed by subtracting each element included in the second input data from each element included in the first input data.

[0089] When the third input data from which the low-frequency noise is removed is generated, the second de-noising unit 162 can insert zero in a blank position generated by moving the first input data. The third input data from which the low-frequency noise is removed can be provided to the cross-correlation unit 132 as the first digital signal from which the noise is removed. In other words, the second de-noising unit 162 can remove low-frequency noise of the interpolated first digital signal and output the first digital signal from which the noise is removed to the cross-correlation unit 132 of the controller 130. Also, the second de-noising unit 162 can remove low-frequency noise of the interpolated second digital signal and output the second digital signal from which the noise is removed to the cross-correlation unit 132 of the controller 130.

[0090] The first digital signal from which the noise is removed can be referred to as a transmission signal from which the noise is removed, and the second digital signal from which the noise is removed can be referred to as a reception signal from which the noise is removed.

[0091] Since the object detection device 100 of the disclosure not only removes high-frequency noise of the transmission signal and the reception signal but also removes low-frequency noise thereof, the detection distance is significantly increased.

[0092] The controller 130 can generate a cross-correlation signal indicating a correlation between the noise-removed transmission signal and the noise-removed reception signal. Also, the controller 130 can detect at least one peak from the cross-correlation signal. Also, the controller 130 can determine a quality of the cross-correlation signal based on the peak of the cross-correlation signal. Also, the controller 130 can acquire a 3D image of the object based on the quality of the cross-correlation signal. To this end, the controller 130 can include a cross-correlation unit 132, a peak detection unit 133, a quality calculation unit 134, and a point cloud generation unit 135.

[0093] The cross-correlation unit 132 can receive the noise-removed transmission signal and the noise-removed reception signal from the filter unit 160.

[0094] The cross-correlation unit 132 can generate a cross-correlation signal between the noise-removed transmission signal and the noise-removed reception signal. To this end, the cross-correlation unit 132 can include a correlator. In one embodiment, the cross-correlation unit 132 can generate the cross-correlation signal by Equation 1 below.

[0095] [Equation 1]

[0096]

[0097] In Equation 1, S1 can denote the transmission signal, S2 can denote the reception signal, and S3 can denote the cross-correlation signal.

[0098] The cross-correlation unit 132 can output the cross-correlation signal to the peak detection unit 133.

[0099] The peak detection unit 133 can detect at least one peak from the cross-correlation signal. The peak detection unit 133 can detect a first peak having a maximum absolute value among peaks of the cross-correlation signal. Also, the peak detection unit 133 can detect a second peak having a maximum absolute value among remaining peaks other than the first peak. The peak detection unit 133 can output the first peak and the second peak to the quality calculation unit 134.

[0100] The quality calculation unit 134 can calculate a quality of the cross-correlation signal based on the absolute value of the first peak and the absolute value of the second peak.

[0101] The quality calculation unit 134 can provide quality information of the cross-correlation signal to the point cloud generation unit 135.

[0102] The point cloud generation unit 135 can generate a point cloud based on the quality information of the cross-correlation signal. Also, the point cloud generation unit 135 can acquire a 3D image of the object based on the point cloud.

[0103] Specifically, the point cloud generation unit 135 can generate a point cloud based on the cross-correlation signals having a predetermined reference quality or greater quality. For example, the reference quality can be set to 2, but is not limited thereto.

[0104] The point cloud generation unit 135 can calculate a transmission time of the transmission signal and a reception time of the reception signal based on the first peak value of the cross-correlation signal having the reference quality or greater quality. Also, the point cloud generation unit 135 can calculate a time of flight of light output from the transmitter 110 based on the transmission time and the reception time. Also, the point cloud generation unit 135 can calculate a distance to the object based on the time of flight of the light. Also, the point cloud generation unit 135 can generate a 3D point cloud based on the distance information to the object.

[0105] The point cloud generation unit 135 can ignore the peak value of the cross-correlation signal lower than the reference quality. Also, the point cloud generation unit 135 can map the maximum detection distance information to the point cloud corresponding to the cross-correlation signal having a quality less than the reference quality. For example, when the maximum detection distance of the object detection device 100 is 200 m, the point cloud generation unit 135 can uniformly store the distance information of 200 m in the point cloud corresponding to the cross-correlation signal having a quality less than the reference quality.

[0106] Since the object detection device 100 according to an example embodiment generates a point cloud by considering the quality of the cross-correlation signal, it has an effect of significantly reducing the possibility of false detection of an object.

[0107] In Figure 2 , although it is depicted that the converter 140, the interpolation unit 150, and the filter unit 160 are shown in a configuration separate from the controller 130, according to an example embodiment, the converter 140, the interpolation unit 150, and the filter unit 160 can be included in the controller 130 as a partial configuration of the controller 130. Also, according to an example embodiment, the cross-correlation unit 132, the peak detection unit 133, the quality calculation unit 134, and the point cloud generation unit 135 in the controller 130 can operate as an additional configuration separate from the control unit 130.

[0108] Figure 3 is a diagram for explaining a method of interpolating a transmission signal and a reception signal according to an example embodiment.

[0109] Referring to Figure 3 , the converter 140 can convert the transmission signal and the reception signal into vector data in a column vector or a row vector form and can output the vector data. In Figure 3Hereinafter, the converter 140 outputs the transmission signal and the reception signal as vector data 310 in the form of a row vector, but the following description also applies to vector data in the form of a column vector. At this time, the left direction and the right direction of the row vector can correspond to the upper direction and the lower direction of the column vector, respectively.

[0110] The interpolation unit 150 can interpolate the vector data 310 by predicting elements between sampling periods. Since the quantized values of the transmission signal and the reception signal based on the sampling periods are stored in each element of the vector data 310, the meaning that the interpolation unit 150 interpolates between the sampling periods can be the same as the meaning that the interpolation unit 150 interpolates between each element included in the vector data 310.

[0111] The interpolation unit 150 can interpolate between a first element E1 and a second element E2 of the vector data 310.

[0112] Specifically, the interpolation unit 150 can calculate an average of a first element E1 value and a second element E2 value. The interpolation unit 150 can insert the average of the first element E1 value and the second element E2 value between the first element E1 and the second element E2 as a first interpolated element E1'. Also, the interpolation unit 150 can insert an average of a second element E2 value and a third element E3 value between the second element E2 and the third element E3 as a second interpolated element E2'.

[0113] The interpolation unit 150 can interpolate between each element included in the vector data 310 by the above-described method. However, the interpolation between elements by the average is only an example, and is not limited thereto. In other words, the vector data 310 can be interpolated using various interpolation methods.

[0114] In Figure 3 , vector data 320 interpolated by any one of various interpolation methods is shown.

[0115] As Figure 3 indicated, the object detection device 100 according to the embodiment has an effect of significantly increasing the sampling rate of the transmission signal and the reception signal without using an analog-to-digital converter of a high sampling rate.

[0116] The interpolation unit 150 according to an example embodiment is disposed between the converter 140 and the filter unit 160 to interpolate the transmission signal before high-frequency noise is removed and the reception signal before high-frequency noise is removed.

[0117] When the interpolation unit 150 interpolates the transmission signal before high-frequency noise is removed and the reception signal before high-frequency noise is removed, there is an effect of significantly improving the distance resolution of the object detection device 100. This will be described with reference to Figures 4-5The effect of increasing the distance resolution according to the interpolation of the transmission signal and the reception signal is described in more detail.

[0118] Figure 4 is a graph for explaining the effect of increasing the distance resolution according to the interpolation of the transmission signal and the reception signal. Figure 5 is a graph for explaining the effect of increasing the distance resolution according to the interpolation of the transmission signal and the reception signal.

[0119] For convenience of explanation, in Figures 4-5 , the effect of increasing the distance resolution is described using the reception signal, but the following description can also be applied to the transmission signal.

[0120] Referring to Figure 4 , the reception signal 410 is depicted in Figure 4 . When the object detection device 100 interpolates the reception signal after removing the high-frequency noise, the object detection device 100 can accumulate the reception signal 410. The first accumulated signal 420 accumulated by the object detection device 100 can be interpolated. Since the object detection device 100 interpolates the first accumulated signal 420 in the state of accumulating the reception signal 410, the interpolated first accumulated signal 430 can include a plurality of peaks Pk1 and Pk2.

[0121] As shown in Figure 4 , when the object detection device 100 interpolates the accumulated first accumulated signal 420 after accumulating the reception signal 410, the sampling rate can increase, but the distance resolution can not increase. On the other hand, the object detection device 100 according to the embodiment can increase the distance resolution by interpolating the reception signal before removing the high-frequency noise.

[0122] Specifically, the object detection device 100 according to the example embodiment can resample, upsample, or interpolate each reception signal 411, 412, 413, and 414. Also, the object detection device 100 can resample, upsample, or accumulate each interpolated reception signal 415, 416, 417, and 418. When the object detection device 100 interpolates each reception signal 411, 412, 413, and 414, and then accumulates the interpolated reception signals 415, 416, 417, and 418, one peak Pk3 can be included in the second accumulated signal 440.

[0123] As shown in Figure 4As shown, when the object detection device 100 interpolates each of the received signals 411, 412, 413, and 414 and then accumulates the interpolated received signals 415, 416, 417, and 418, the distance to the object can be further finely divided. In other words, since the object detection device 100 according to the example embodiment interpolates each of the received signals 411, 412, 413, and 414 and then accumulates the interpolated received signals 415, 416, 417, and 418, the distance resolution can be significantly increased.

[0124] Figure 5 A first distance resolution graph 510 and a second distance resolution graph 520 are shown, in which the first distance resolution graph 510 represents a case in which the object detection device 100 accumulates the received signals 410 and then interpolates the accumulated first accumulated signal 420, and the second distance resolution graph 520 represents a case in which the object detection device 100 interpolates each of the received signals 411, 412, 413, and 414 and then accumulates the interpolated respective received signals 415, 416, 417, and 418.

[0125] As shown, when the object detection device 100 interpolates each of the received signals 411, 412, 413, and 414 and then accumulates the interpolated received signals 415, 416, 417, and 418, the distance to the object can be further finely divided. In other words, since the object detection device 100 according to the example embodiment interpolates each of the received signals 411, 412, 413, and 414 and then accumulates the interpolated received signals 415, 416, 417, and 418, the distance resolution can be significantly increased. Figure 5 As shown, when the object detection device 100 interpolates each of the received signals 411, 412, 413, and 414 and then accumulates the interpolated received signals 415, 416, 417, and 418, the distance to the object can be further finely divided. In other words, since the object detection device 100 according to the example embodiment interpolates each of the received signals 411, 412, 413, and 414 and then accumulates the interpolated received signals 415, 416, 417, and 418, the distance resolution can be significantly increased.

[0126] Figure 6 is a graph for explaining a method of removing low-frequency noise of a transmitted signal and a received signal according to an example embodiment.

[0127] Referring to Figure 6 , the second denoising unit 162 can receive the vector data from which the high-frequency noise is removed as the first input data 610.

[0128] The second denoising unit 162 can generate the second input data 620 by moving each element included in the first input data 610 in a predetermined direction and by a predetermined size. When moving the elements, the second denoising unit 162 can delete elements outside a column address range. Accordingly, the length of the second input data 620 can be less than the length of the first input data 610.

[0129] Specifically, when the first input data 610 is in the form of a row vector, the second de-noising unit 162 can perform element shifting to the left or to the right on the first input data 610. In this case, the left can denote a direction in which a column address of the row vector is decreased, and the right can denote a direction in which the column address of the row vector is increased. Also, the second de-noising unit 162 can perform element shifting of 1 to 500 elements on the first input data 610.

[0130] For example, as shown in FIG. 6B, the second de-noising unit 162 can generate the second input data 620 by shifting the first input data 610 by 2 elements in the left direction. At this time, the third element E3 included in the first input data 610 can be set in the first column of the second input data 620 by decreasing the column address by 2. Also, the first element E1 and the second element E2 outside the range of the column address can be deleted. Figure 6

[0131] The second de-noising unit 162 can generate the third input data 630 from which low-frequency noise is removed, by subtracting each element included in the second input data 620 from each element included in the first input data 610 based on the column address.

[0132] On the other hand, when the third input data 630 is generated, the second de-noising unit 162 can punch in zeros in the blank position 631 of the third input data 630 generated by shifting the first input data 610.

[0133] Specifically, the second de-noising unit 162 can subtract elements set at column addresses corresponding to each other from among the first input data 610 and the second input data 620.

[0134] The second de-noising unit 162 can store the subtracted elements in the third input data 630. At this time, the second de-noising unit 162 can store the subtracted elements in the third input data 630 at positions corresponding to the column addresses of the first input data 610 and the second input data 620. For example, when elements set at position (1, 1) among elements included in the first input data 610 and the second input data 620 are subtracted, the second de-noising unit 162 can store the subtracted elements at position (1, 1) of the third input data 630.

[0135] When there is no element corresponding to an element selected from the first input data 610 in the second input data 620, the second de-noising unit 162 can insert a zero in the third input data 630 at a position corresponding to the column address of the first input data 610. For example, in Figure 6 ​In the second denoising unit 162, since there is no second input data corresponding to positions (1, 9) and (1, 10) of the first input data 610, the second denoising unit 162 can insert zeros in positions (1, 9) and (1, 10) of the third input data 630.

[0136] Figure 7 is a graph for explaining an effect when zeros are inserted in the blank positions of the third input data.

[0137] Figure 7 shows a first reception signal 711 generated when no zero is inserted in the blank positions of the third input data, a first cross-correlation signal 712 between the first reception signal 711 and the transmission signal, and a first detection distance graph 713 using the first cross-correlation signal 712. Also, Figure 7 shows a second reception signal 714 generated when zeros are inserted in the blank positions of the third input data, a second cross-correlation signal 715 between the second reception signal 714 and the transmission signal, and a second detection distance graph 716 using the second cross-correlation signal 715.

[0138] As Figure 7 can be seen, when zeros are inserted in the blank positions of the third input data, the uncomputable region 714a of the second reception signal 714 is stabilized. Thus, it can be seen that the uncomputable region 715a of the second cross-correlation signal 715 is also stabilized.

[0139] As in the first detection distance graph 713 and the second detection distance graph 716, the first cross-correlation signal 712 is unstable, and thus, the detection distance rapidly changes between 0 m and 80 m. However, since the second cross-correlation signal 715 is stable, the change in the detection distance of the second detection distance graph 716 is significantly smaller than the change in the detection distance of the first detection distance graph 713.

[0140] Figures 8A-8B and Figure 9 is a graph for explaining an effect of increasing the detection distance according to removing the low-frequency noise of the transmission signal and the reception signal.

[0141] More specifically, Figure 8A is a graph for explaining an effect of the shape of the cross-correlation signal 814 and noise when the object detection device 100 generates the cross-correlation signal 814 between the transmission signal 811 and the reception signal 812 after removing only the high-frequency noise from the reception signal 812. Also, Figure 8B is a graph for explaining an effect of the shape of the cross-correlation signal 820 and noise when the object detection device 100 generates the cross-correlation signal 820 between the transmission signal 815 and the reception signal 817 after removing the high-frequency noise and the low-frequency noise from the reception signal 816.

[0142] Referring to Figure 8A , the object detection device 100 can receive a transmission signal 811 and a reception signal 812. Since the reception signal 812 is a signal reflected from an object, the size of the reception signal 812 can be smaller than that of the transmission signal 811. Also, noise can be included in the reception signal 812.

[0143] The object detection device 100 can output a reception signal 813 in which high-frequency noise is removed, by accumulating the reception signal 812 for a predetermined time and then calculating the average of the accumulated reception signal. The reception signal 812 can still include a low-frequency noise component.

[0144] The object detection device 100 can generate a cross-correlation signal 814 between the transmission signal 811 and the reception signal 813 in which high-frequency noise is removed. The object detection device 100 can calculate the ToF of light based on the peak of the cross-correlation signal 814 and can generate a point cloud based on the ToF of light.

[0145] As shown in FIG. 11, when the object detection device 100 generates the cross-correlation signal 814 after removing only high-frequency noise, since the cross-correlation signal 814 includes low-frequency noise and the definition of the cross-correlation signal 814 is low, the object detection device 100 can not detect an accurate peak in the cross-correlation signal 814. Also, detecting an inaccurate peak results in a rapid change in a detection distance. Figure 8A

[0146] On the other hand, the object detection device 100 according to an example embodiment generates a cross-correlation signal after removing not only high-frequency noise but also low-frequency noise, thereby improving a detection distance.

[0147] Specifically, in FIG. 12, the object detection device 100 according to an example embodiment can receive a transmission signal 815 and a reception signal 816. Figure 8B The transmission signal 815 of FIG. 12 can correspond to the transmission signal 811 of FIG. 10, Figure 8B The reception signal 816 of FIG. 12 can correspond to the reception signal 812 of FIG. 10. Figure 8A Figure 8B Figure 8A

[0148] The object detection device 100 can output a reception signal 818 in which high-frequency noise is removed, by accumulating the reception signal 817 for a predetermined time and then calculating the average of the accumulated reception signal.

[0149] ​​​​Also, the object detection device 100 can output the transmission signal 816 from which the low-frequency noise is removed and the reception signal 819 from which the low-frequency noise is removed by subtracting elements corresponding to each other after element shifting is performed on each of the transmission signal 815 and the reception signal 818.

[0150] The object detection device 100 can generate a cross-correlation signal 820 between the transmission signal 816 from which the low-frequency noise is removed and the reception signal 819 from which the low-frequency noise is removed. The object detection device 100 can calculate the ToF of light based on a peak value of the cross-correlation signal 820 and can generate a point cloud based on the ToF of light.

[0151] As shown in FIG. 11, Figure 8B It can be seen that noise of the cross-correlation signal 820 is mostly removed. In addition, since the cross-correlation signal 820 is generated based on the transmission signal 816 from which the low-frequency noise is removed and the reception signal 819 from which the low-frequency noise is removed, the definition of the cross-correlation signal 820 can be high. Accordingly, the object detection device 100 can more accurately detect a peak value from the cross-correlation signal 820, and thus, the detection distance of the object detection device 100 can be improved.

[0152] Figure 9 A third detection distance graph 911 and a fourth detection distance graph 912 are shown, in which the third detection distance graph 911 indicates a case in which the object detection device 100 generates a cross-correlation signal 911 after removing only high-frequency noise, and the fourth detection distance graph 912 indicates a case in which the object detection device 100 generates a cross-correlation signal 913 after removing both high-frequency noise and low-frequency noise. In Figure 9 In the above,

[0153] As shown in FIG. 11, Figure 9 When the object detection device 100 generates the cross-correlation signal 911 after removing only high-frequency noise, since a correct peak position can not be detected in the cross-correlation signal 911, the detection distance can rapidly change between 0 m and 150 m. However, when the object detection device 100 generates the cross-correlation signal 913 after removing both high-frequency noise and low-frequency noise, since a correct peak position can be detected in the cross-correlation signal 913, the change in the detection distance is significantly reduced.

[0154] Figure 10 is a graph for explaining a method of calculating a quality of a cross-correlation signal according to an example embodiment.

[0155] In the above, Figure 10 In the above,

[0156] Referring to Figure 10, the peak detection unit 133 can detect a first peak P1 having a largest absolute value among peaks of the cross-correlation signal S3. Also, the peak detection unit 133 can detect a second peak P2 having a largest absolute value among the remaining peaks other than the first peak P1. The peak detection unit 133 can output the first peak P1 and the second peak P2 to the quality calculation unit 134.

[0157] The quality calculation unit 134 can calculate the quality of the cross-correlation signal based on an absolute value of the first peak P1 and an absolute value of the second peak P2.

[0158] In one embodiment, the quality calculation unit 134 can calculate the quality of the cross-correlation signal by using any one of the following Equations 2 to 5.

[0159] [Equation 2]

[0160] S Qos = ||P1 - |P2||

[0161] [Equation 3]

[0162]

[0163] [Equation 4]

[0164] S QoS = log(||P1 - |P2||)

[0165] [Equation 5]

[0166]

[0167] In Equations 2 to 5, S QoS may be the quality of the cross-correlation signal, P1may be the first peak, and P2may be the second peak.

[0168] The point cloud generation unit 135 can generate a point cloud based on the cross-correlation signal having a predetermined reference quality or greater. In other words, the point cloud generation unit 135 can ignore the cross-correlation signal having a small difference, ratio, etc., between the first peak P1 and the second peak P2.

[0169] When the point cloud generation unit 135 generates a point cloud based on the cross-correlation signal having a predetermined reference quality or greater, the possibility of error detection can be significantly reduced.

[0170] Figure 11 is a graph for explaining an effect of preventing error detection according to quality calculation of a cross-correlation signal.

[0171] Referring to Figure 11, shows an original image, a cross-correlation signal having a predetermined quality or more, and an object image from which noise is removed.

[0172] As shown in Figure 11 When the quality of the cross-correlation signal is determined and a 3D image is generated based on the determined quality, the object detection device 100 can acquire the same 3D image as the original image.

[0173] Figure 12 is a flowchart of an object detection method according to an example embodiment.

[0174] Referring to Figure 12 In operation S1210, the object detection device 100 can receive a transmission signal transmitted toward an object and a reception signal reflected from the object. At this time, the transmission signal can mean a portion of the transmission signal transmitted toward the object.

[0175] In operation S1220, the object detection device 100 can convert each of the transmission signal and the reception signal into a digital signal according to a predetermined sampling period.

[0176] The object detection device 100 can output the digital signal as vector data in the form of a column vector or a row vector. The vector data can mean an array in the form of a column vector or a row vector including a set of elements. Quantized values of the transmission signal and the reception signal can be stored in each element.

[0177] In operation S1230, the object detection device 100 can interpolate between sampling periods.

[0178] The object detection device 100 can interpolate between each element included in the vector data by predicting elements between sampling periods. For example, the object detection device 100 can interpolate the transmission signal and the reception signal by using at least one of a linear interpolation method, a polynomial interpolation method, a spline interpolation method, an exponential interpolation method, a logarithmic linear interpolation method, a Lagrange interpolation method, a Newton interpolation method, and a bilinear interpolation method. However, the above-mentioned interpolation methods are only examples of interpolation methods, and various interpolation methods can be used to interpolate the transmission signal and the reception signal.

[0179] In operation S1240, the object detection device 100 can remove noise from each of the interpolated transmission signal and the interpolated reception signal.

[0180] The object detection device 100 can remove high-frequency noise from each of the interpolated transmission signal and the interpolated reception signal, and then remove low-frequency noise from each of the transmission signal from which the high-frequency noise is removed and the reception signal from which the high-frequency noise is removed.

[0181] Reference will be made to Figure 14 and Figure 15The method of removing noise by the object detection device 100 will be described in more detail.

[0182] In operation S1250, the object detection device 100 can generate a cross-correlation signal between the noise-removed transmission signal and the noise-removed reception signal.

[0183] The object detection device 100 can generate a cross-correlation signal between the noise-removed transmission signal and the noise-removed reception signal by using Equation 1 described above, but is not limited thereto.

[0184] In operation S1260, the object detection device 100 can acquire a 3D image of the object based on a peak value of the cross-correlation signal.

[0185] Figure 13 is a flowchart for explaining a method of removing high-frequency noise of Figure 12 .

[0186] Referring to Figure 13 , in operation S1310, the object detection device 100 can accumulate each element value included in the interpolated vector data for a predetermined time.

[0187] In operation S1320, the object detection device 100 can remove high-frequency noise by outputting an average value of the accumulated elements.

[0188] According to an example embodiment, each operation of Figure 13 can be performed only on the reception signal.

[0189] Figure 14 is a flowchart for explaining a method of removing low-frequency noise of Figure 12 .

[0190] Referring to Figure 14 , in operation S1410, the object detection device 100 can receive the vector data from which the high-frequency noise is removed as first input data.

[0191] In operation S1420, the object detection device 100 can generate second input data by moving each element included in the first input data in a predetermined direction and by a predetermined size.

[0192] The predetermined direction can be any one of left, right, up, and down directions. When the vector data is a row vector, the predetermined direction can be set to left or right. At this time, the left side can indicate a direction in which the column address of the vector data decreases, and the right side can indicate a direction in which the column address of the vector data increases. Also, when the vector data is a column vector, the predetermined direction can be set to the upper side or the lower side. At this time, the upper side can indicate a direction in which the row address of the vector data decreases, and the lower side can indicate a direction in which the row address of the vector data increases. The predetermined size can be set to 500 elements.

[0193] When the elements are moved, the object detection device 100 can delete elements outside the column address range. Accordingly, the length of the second input data can be less than the length of the first input data.

[0194] In operation S1430, the object detection device 100 can output third input data from which low-frequency noise is removed, by subtracting each element included in the second input data from each element included in the first input data.

[0195] The object detection device 100 can subtract elements arranged at column addresses corresponding to each other in the first input data and the second input data.

[0196] The object detection device 100 can store the subtracted elements in the third input data. At this time, the object detection device 100 can store the subtracted elements in positions of the third input data corresponding to the column addresses of the first input data and the second input data.

[0197] On the other hand, when there is no element corresponding to the element selected from the first input data in the second input data, the object detection device 100 can insert zero at a position of the third input data corresponding to the column address of the first input data.

[0198] Figure 15 is a flowchart for explaining a method of acquiring a 3D image. Figure 12

[0199] Referring to Figure 15 In operation S1510, the object detection device 100 can detect at least one peak value in the cross-correlation signal.

[0200] The object detection device 100 can detect a first peak value having a greatest absolute value among the peak values of the cross-correlation signal. Also, the object detection device 100 can detect a second peak value having a greatest absolute value among the remaining peak values other than the first peak value.

[0201] In operation S1520, the object detection device 100 can determine the quality of the cross-correlation signal based on the peak value.

[0202] ​The object detection device 100 can determine the quality of the cross-correlation signal based on the absolute value of the first peak value and the absolute value of the second peak value. The object detection device 100 can determine the quality of the cross-correlation signal according to Equations 2 to 5.

[0203] In operation S1530, the object detection device 100 can generate a point cloud based on the cross-correlation signal having the predetermined quality or greater.

[0204] The object detection device 100 can calculate a transmission time of the transmission signal and a reception time of the reception signal based on the first peak value of the cross-correlation signal having the reference quality or greater. Also, the object detection device 100 can calculate the ToF of the light output from the transmitter 110 based on the transmission time and the reception time. Also, the object detection device 100 can calculate the distance to the object based on the ToF of the light. Also, the object detection device 100 can generate a 3D point cloud based on the distance information to the object.

[0205] The object detection device 100 can ignore the peak value of the cross-correlation signal having a quality less than the reference quality. In addition, the object detection device 100 can map the maximum detection distance information to the point cloud corresponding to the cross-correlation signal having a quality less than the reference quality.

[0206] In operation S1540, the object detection device 100 can acquire a 3D image based on the generated point cloud.

[0207] Although not limited thereto, the example embodiments can be embodied as computer readable code on a computer readable recording medium. The computer readable recording medium is any data storage device that can store data which can be thereafter read by a computer system. Examples of the computer readable recording medium include read-only memory (ROM), random-access memory (RAM), CD-ROMs, magnetic tapes, floppy disks, and optical data storage devices. The computer readable recording medium can also be distributed over network-coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. Also, the example embodiments can be written as computer programs and can be implemented in general-use or special-purpose digital computers that execute the programs as computer programs written in a computer language by using a computer readable recording medium. Furthermore, it should be understood that one or more units of the above-described apparatus and device can include circuitry, a processor, a microprocessor, etc., and can execute a computer program stored in a computer readable medium.

[0208] The foregoing exemplary embodiments are merely illustrative, and should not be construed as limiting. The present teachings can be readily applied to other types of apparatuses. Also, the description of the exemplary embodiments is intended to be illustrative, and not to limit the scope of the claims, as many alternatives, modifications and variations will be apparent to those skilled in the art.

Claims

1. An object detection apparatus comprising: a converter configured to convert a transmission signal transmitted toward an object and a reception signal reflected from the object into digital transmission and reception signals according to a predetermined sampling period; and at least one processor configured to: interpolate between elements of the digital transmission and reception signals having the predetermined sampling period to obtain interpolated transmission and reception signals; remove noise from each of the interpolated transmission and reception signals; generate a cross-correlation signal between the interpolated transmission signal with the noise removed and the interpolated reception signal with the noise removed; and obtain a three-dimensional (3D) image of the object based on at least one peak of the cross-correlation signal, wherein the at least one processor is further configured to: detect a first peak having a largest absolute value among the at least one peak of the cross-correlation signal; detect a second peak having a second largest absolute value among the at least one peak of the cross-correlation signal; determine a quality of the cross-correlation signal based on the absolute value of the first peak and the absolute value of the second peak; generate a point cloud based on cross-correlation signals having a quality greater than or equal to a predetermined reference quality; and obtain the 3D image of the object based on the generated point cloud. the converter is configured to output each of the transmission and reception signals by converting each of the transmission and reception signals into vector data in a column vector or a row vector form.

2. The object detection device according to claim 1, wherein the at least one processor is further configured to interpolate between each element included in the vector data.

3. The object detection device of claim 2, wherein, the at least one processor is further configured to remove high frequency noise by accumulating each element included in the interpolated vector data over a predetermined time and outputting an average value of the accumulated elements, and 4. The object detection device according to claim 3, wherein wherein the high frequency noise is a portion of noise having a frequency higher than a predetermined upper threshold. the at least one processor is further configured to:

5. The object detection device of claim 4, wherein, receive the vector data with the high frequency noise removed as first input data; generate second input data by moving each first element included in the first input data by a predetermined size in a predetermined direction; and output third input data with low frequency noise removed by subtracting each second element included in the second input data from each first element included in the first input data, and wherein the low frequency noise is a portion of noise having a frequency lower than a predetermined lower threshold. the at least one processor is further configured to insert zero in the third input data when there is no second element corresponding to any first element of the first input data in the second input data.

6. The object detection device of claim 5, wherein, 7.A method for object detection, the method comprising: converting a transmission signal transmitted toward an object and a reception signal reflected from the object into digital transmission and reception signals according to a predetermined sampling period; interpolating between elements of the digital transmission and reception signals having the predetermined sampling period to obtain interpolated transmission and reception signals; ​ removing noise from each of the interpolated transmission signal and the interpolated reception signal; generating a cross-correlation signal between the noise-removed interpolated transmission signal and the noise-removed interpolated reception signal; and acquiring a three-dimensional (3D) image of the object based on at least one peak of the cross-correlation signal, wherein the acquiring of the 3D image of the object includes: detecting a first peak having a largest absolute value among the at least one peak of the cross-correlation signal; detecting a second peak having a second largest absolute value among the at least one peak of the cross-correlation signal; determining a quality of the cross-correlation signal based on the absolute value of the first peak and the absolute value of the second peak; generating a point cloud based on the cross-correlation signal having a quality greater than or equal to a predetermined reference quality; and acquiring the 3D image of the object based on the generated point cloud.

8. The method of claim 7, wherein, The converting includes converting each of the transmission signal and the reception signal into vector data in a column vector or a row vector form, and The interpolating includes interpolating between each element included in the vector data.

9. The method of claim 8, wherein, The removing includes: removing high-frequency noise by accumulating each element included in the interpolated vector data over a predetermined time and outputting an average value of the accumulated elements; receiving the vector data from which the high-frequency noise is removed as first input data; generating second input data by moving each first element included in the first input data by a predetermined size in a predetermined direction; and outputting third input data from which low-frequency noise is removed by subtracting each second element included in the second input data from each first element included in the first input data, and wherein the high-frequency noise is a part of noise having a frequency higher than a predetermined upper threshold value.

10. The method of claim 9, wherein, The outputting of the third input data includes inserting zero in the third input data when there is no second element corresponding to any first element of the first input data in the second input data.

Images