Distance measurement method and equipment based on TOF camera and binocular vision data fusion

Abstract

The present invention relates to a distance measurement method and device based on the fusion of TOF camera and binocular vision data, belonging to the field of fusion distance measurement technology. In order to solve the problem of low distance measurement accuracy caused by system uncertainty, environmental interference and invalid data in the existing distance measurement method, the present invention first uses a Gaussian model to process each set of data, eliminates distance values ​​with excessive deviations, selects high-probability distance values ​​and calculates the average value to obtain the optimal distance measurement value; the measurement data uses the binocular vision distance measurement average value as the final distance measurement value in the close-range segment, and the TOF camera distance measurement average value as the final distance measurement value in the long-range segment; an adaptive weighted fusion method is used in the intermediate distance segment, and different weights are assigned to the TOF distance measurement average value and the binocular vision distance measurement average value. The distance measurement average values ​​of the TOF camera and the binocular vision are multiplied by their respective weights and summed as the final distance measurement value of the intermediate distance segment.

Description

Technical Field

[0001] The present invention belongs to the technical field of fusion ranging and relates to a fusion ranging method, a storage medium and a device. Background Art

[0002] With the advancement of technology and scientific development, fields such as autonomous driving, robot vision, and equipment docking have put forward higher requirements for distance accuracy. The accuracy and efficiency of sensor detection results will directly affect the safety and reliability of the project.

[0003] In the field of optics, commonly used distance measurement methods include time-of-flight (TOF), structured light, and binocular vision. Structured light has a short measurement range and can only be used indoors. Natural light in outdoor environments can overwhelm the coded light, limiting its applicability. TOF cameras directly measure light's time of flight and are less affected by changes in lighting and texture, allowing for measurement at greater distances. Binocular vision uses the principle of eye imaging to measure distance, achieving high accuracy at close ranges. Both TOF cameras and binocular vision have their own advantages and disadvantages in different measurement ranges and scenarios.

[0004] Taking into account the limitations of single-sensor measurement methods, multi-sensor information fusion technology can be used to improve the distance measurement accuracy by using two sensors, TOF camera and binocular vision, to measure the distance of the same target parameter. At the same time, since a single data fusion method has certain limitations, integrating the advantages of two or more data fusion methods can effectively reduce the negative impact of system uncertainty, environmental interference and invalid data on state estimation. For distance measurement, this will lead to relatively low distance measurement accuracy. Most of the existing depth information fusion methods related to TOF cameras require the generation of a TOF camera disparity map and then pixel-level processing. This method not only needs to be improved in terms of effectiveness, but the pixel-level processing also leads to slow processing speed and low efficiency. Summary of the Invention

[0005] The present invention aims to solve the problems of low ranging accuracy due to system uncertainty, environmental interference and invalid data in existing distance measurement methods, as well as ranging errors caused by system errors in a single ranging system.

[0006] A distance measurement method based on the fusion of TOF camera and binocular vision data includes the following steps:

[0007] S1. Obtain the distance data of the target to be measured by the TOF camera and the binocular vision ranging platform. The obtained data corresponds to n measurements, that is, the TOF camera and the binocular vision camera each obtain n distance values ​​D i , i=1,2,…,n;

[0008] The TOF camera and binocular vision ranging platform includes a TOF camera and a binocular vision camera. The binocular vision camera includes a binocular vision right camera and a binocular vision left camera. The centers of the TOF camera and the left and right binocular vision cameras are in a straight line, and this straight line and the line connecting the platform and the target are perpendicular to each other.

[0009] S2. Perform the following processing for TOF camera and binocular vision respectively:

[0010] S2.1. Determine the Gaussian distribution function of the random measurement value d:

[0011]

[0012] Among them, σ is the standard deviation, d is the n distance value D i The random measurement value in, μ is the mathematical expectation, σ 2 is n distance values ​​D i The corresponding variance;

[0013] S2.2: Determine the range of possible values ​​and the critical value of the Gaussian distribution function:

[0014]

[0015] Where u is the lower critical value of the Gaussian distribution function;

[0016] When the value of the Gaussian distribution function is greater than u, the measured value is considered to have a high probability of occurrence; m is the n distance value D i The corresponding average value;

[0017] S2.3: Select the distance value X retained after Gaussian model screening from the initial distance value i , the number is r; get the optimal value of ranging:

[0018] where X i is the i-th value that meets the requirements, i = 1, 2, ..., r, r is the number of values ​​that meet the requirements;

[0019] S3, determining whether any of the optimal distances obtained by the TOF and binocular methods is between the first distance threshold and the second distance threshold; if so, executing S4;

[0020] S4. The average value of TOF distance measurement and the average value of binocular vision distance measurement are fused according to the weights to obtain the final distance measurement value:

[0021] S4.1. Let the TOF camera be denoted as p, and the data after the Gaussian model is denoted as X p , the mean is

[0022] The binocular vision camera is recorded as q, and the data after the Gaussian model is recorded as X q , the mean is

[0023] S4.2. Assume that the actual distance is X ture , the observation error between the average value of TOF ranging and binocular vision ranging is recorded as V p and V q , then X p =X ture +V p and X q =X ture +V q Observation error V p and V q Considered as zero-mean stationary noise;

[0024] The data measured by camera p has a variance of σ 2 =E(V p 2 ), where E(·) is the expectation;

[0025] S4.3. Since both cameras have r data after being processed by the Gaussian model, the cross-correlation function R between camera p and camera q is obtained. pq , and the autocorrelation function R of camera p pp :

[0026]

[0027]

[0028] S4.4. Determine the variance of camera p And the variance of camera q:

[0029]

[0030]

[0031] S4.5. Determine the weighting factor W of camera p and camera q p and W q :

[0032]

[0033]

[0034] S4.5. Using the weighting factor W p and W q After fusion, the estimated value of the fused distance X is as follows:

[0035]

[0036] in, and It is the mean value of the data of TOF camera and binocular vision after being filtered by Gaussian model.

[0037] Furthermore, the process of determining the first distance threshold and the second distance threshold is as follows:

[0038] A1. Obtain the distance data of the target to be measured by the TOF camera and the binocular vision ranging platform. The obtained data corresponds to L / k measurement positions, and each position corresponds to n measurement data;

[0039] L / k measurement positions: Within the range L of the target, measurement positions separated by a distance of k are placed close to the target and the distance to the target is measured in sequence. Each measurement position is measured n times, for a total of L / k measurement positions.

[0040] A2. Process the TOF camera and binocular vision separately, and use the Gaussian model to process each set of data to obtain the optimal distance measurement value;

[0041] A3. Based on L / k measurement positions, determine the optimal distance and the corresponding distance error, and obtain the error-distance curves of the TOF and binocular vision methods within the ranging range of 0 to L, respectively; plot the error-distance curves of the TOF and binocular vision methods together, and the intersection of the minimum distance value and the maximum distance value of the two curves is the first distance threshold and the second distance threshold.

[0042] Furthermore, the processing procedures of step A2 are the same as those of step S2.

[0043] Preferably, the first distance threshold and the second distance threshold are 1 m and 3.5 m respectively.

[0044] Furthermore, the method further comprises the following steps:

[0045] Based on the target location, the distance within the first distance threshold is regarded as the short distance segment, and the distance outside the second distance threshold is regarded as the long distance segment;

[0046] In the process of determining in S3 whether any of the optimal distances of the ranging values ​​obtained by the TOF and binocular methods is between the first distance threshold and the second distance threshold, if the optimal distances of the ranging values ​​obtained by the TOF and binocular methods are both in the long-distance segment, the average distance value corresponding to the camera with the smaller distance error between the TOF camera and the binocular vision camera in the long-distance segment is selected as the final ranging value; if the optimal distances of the ranging values ​​obtained by the TOF and binocular methods are both in the short-distance segment, the average distance value corresponding to the camera with the smaller distance error between the TOF camera and the binocular vision camera in the short-distance segment is selected as the final ranging value.

[0047] Furthermore, the distance error of the camera with the smaller distance error in the long-distance segment is determined in the process of determining the ranging optimal value distance and the corresponding distance error in A3.

[0048] Furthermore, the distance error of the camera with the smaller distance error in the close distance segment is determined in the process of determining the optimal distance and the corresponding distance error in A3.

[0049] Furthermore, in the short distance segment, the average distance measurement value corresponding to the binocular vision camera is selected as the final distance measurement value; in the long distance segment, the average distance measurement value corresponding to the TOF camera is selected as the final distance measurement value.

[0050] A computer storage medium stores at least one instruction, which is loaded and executed by a processor to implement a distance measurement method based on the fusion of a TOF camera and binocular vision data.

[0051] A distance measurement device based on the fusion of a TOF camera and binocular vision data, the device comprising a processor and a memory, wherein the memory stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement the distance measurement method based on the fusion of a TOF camera and binocular vision data.

[0052] Beneficial effects:

[0053] The present invention proposes a distance measurement method based on the fusion of TOF camera and binocular vision data. TOF camera and binocular vision ranging have their own advantages and disadvantages in different measurement ranges and measurement scenarios. By combining the two sensors to measure the distance of the same target parameter, their advantages complement each other and can improve the ranging accuracy.

[0054] This paper proposes a data fusion method in which two data sets are first processed using a Gaussian model to eliminate values ​​with large errors. The data are then adaptively weighted and fused to obtain the optimal value by assigning different weights. This integration of the advantages of the two data fusion methods can effectively reduce the negative impact of system uncertainty, environmental interference, and failure data on state estimation.

[0055] Most of the existing depth information fusion methods related to TOF cameras require the generation of a TOF camera disparity map. The present invention does not require the generation of a disparity map for the TOF depth map, nor does it require pixel-level fusion processing. Instead, a new data fusion method is proposed with high algorithm efficiency and fast processing time. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a schematic diagram of a distance measurement method based on the fusion of TOF camera and binocular vision data in the present invention.

[0057] Figure 2 This is the binocular vision calibration flow chart.

[0058] Figure 3 This is a model diagram of the principle of adaptive weighted fusion of TOF camera and binocular vision.

[0059] Figure 4 It is a distance measurement diagram. DETAILED DESCRIPTION

[0060] This invention provides a ranging method based on the fusion of TOF camera and binocular vision data. This method fuses the TOF camera and binocular vision ranging results using two fusion methods: Gaussian model processing and adaptive weighting. This method suppresses the influence of errors, expands the sensor's operating conditions, and improves reliability and accuracy. The invention is further described below with reference to the accompanying drawings and specific embodiments.

[0061] Specific implementation method 1: Combination Figure 1 To explain this embodiment,

[0062] This embodiment is a ranging method based on the fusion of TOF camera and binocular vision data. The process of the ranging method based on the fusion of TOF camera and binocular vision data is ranging, first data optimization, second data optimization, and final estimated value output.

[0063] The distance measurement method based on the fusion of TOF camera and binocular vision data described in this embodiment includes the following steps:

[0064] Step 1: Build a binocular vision ranging platform, calibrate the intrinsic and extrinsic parameters of the left and right binocular vision cameras, and obtain the camera's intrinsic parameters and the transformation relationship between the two cameras.

[0065] like Figure 2 As shown, first prepare a black and white checkerboard image on a plane. The left and right cameras take pictures of the image at the same time. Take multiple groups of pictures, detect the corner points of the image through simulation software, and solve the camera's intrinsic and extrinsic parameters, that is, the intrinsic parameters and relative transformation relationship of the left and right cameras.

[0066] Step 2: Build a TOF camera and binocular vision ranging platform to measure the distance of the target, which is the object to be measured (6); Figure 4 As shown, the TOF camera and binocular vision ranging platform include a camera bracket (1), a camera platform (2), a binocular vision right camera (3), a TOF camera (4), and a binocular vision left camera (5); the binocular vision right camera (3), the binocular vision left camera (5), and the TOF camera (4) are arranged on the camera platform, and their centers are on a straight line, and the binocular vision right camera (3) and the binocular vision left camera (5) are arranged on both sides of the TOF camera (4); and the camera platform (2) is arranged on the camera bracket (1).

[0067] The TOF camera and the centers of the left and right cameras of the binocular vision are on a straight line, and this straight line and the line connecting the platform and the target are perpendicular to each other.

[0068] Within the range L, the TOF camera and binocular vision system are used to measure the target at intervals of k. The target is photographed and measured in sequence to obtain the measured image. There are a total of L / k measurement positions. At each measurement position, the TOF camera and binocular vision system are required to measure n times (n ≥ 20).

[0069] Step 3: In order to measure the distance to a certain point on the target, the measured point is first selected on the target. The distance value of this point can be directly obtained on the TOF depth image. Binocular vision ranging requires the use of the semi-global stereo matching algorithm SGBM. After stereo correction and stereo matching (including four stages: preprocessing, cost calculation, dynamic programming, and post-processing), the disparity map is calculated and the distance value of the measured point is obtained in the disparity map.

[0070] At this time, there are L / k measurement positions in the entire ranging range, so the TOF camera and binocular vision each form L / k groups of data. At each measurement position, each group has n distance values. When measuring at the same ranging position, there is no need to move the ranging platform.

[0071] The purpose of L / k measurement positions is to obtain the error-distance curves of the TOF and binocular vision methods within the 0-L distance range. In actual application, for a certain measurement position, the TOF and binocular vision methods only need to measure at a certain position n times (n ≥ 20).

[0072] Step 4: Process the TOF camera and binocular vision separately, and use the Gaussian model to process each set of data. Eliminate distance values ​​with excessive deviations, that is, distance values ​​that do not meet the requirements of step 4.2. The distance values ​​filtered by the Gaussian model are high-probability distance values. Select the high-probability distance values ​​and calculate the average value to obtain the optimal distance measurement value.

[0073] Taking a set of data from a certain camera as an example, the specific steps are as follows:

[0074] Step 4.1: The Gaussian distribution function of the random measurement value d is:

[0075]

[0076] Among them, σ is the standard deviation, d is the n distance value D i The random measurement value in μ is the mathematical expectation value, m, σ 2 There are n distance values ​​D i The corresponding mean and variance;

[0077] The mean m and variance σ of the measurement data 2 They are:

[0078]

[0079]

[0080] Among them, D i is the i-th initial measurement value, i = 1, 2, ..., n. n is the total number of data;

[0081] Step 4.2: Determine the range of optional values ​​and the critical value of the Gaussian distribution function:

[0082]

[0083] Where u is the lower critical value of the Gaussian distribution function, which is generally selected between 0.6 and 0.8. When the value of the Gaussian distribution function is greater than u, the measured value is considered to have a high probability of occurrence; when the value of the Gaussian distribution function is less than u, the measured value is considered to have a small probability error value and can be discarded;

[0084] Step 4.3: Select the distance value X retained after Gaussian model screening from the initial distance value i , the number is r. Get the optimal value of ranging:

[0085]

[0086] where X i is the i-th value that meets the requirements, i = 1, 2, ..., r, r is the number of values ​​that meet the requirements;

[0087] The Gaussian model solves the impact of errors caused by low-probability events on the overall ranging accuracy during ranging, improving the ranging accuracy and stability of the system.

[0088] Based on L / k measurement positions, the optimal distance and the corresponding distance error are determined, and the error-distance curves of the TOF and binocular vision methods in the range of 0 to L are obtained respectively. The applicable ranges of the TOF camera and binocular vision are determined based on the error-distance curves of the TOF and binocular vision methods:

[0089] The error-distance curves of the TOF and binocular vision methods are plotted together, and the intersection point corresponding to the minimum distance value and the intersection point corresponding to the maximum distance value of the two curves are recorded as the first distance threshold and the second distance threshold; the distance within the first distance threshold is regarded as the short distance segment, the distance outside the second distance threshold is regarded as the long distance segment, and the distance between the first distance threshold and the second distance threshold is regarded as the intermediate distance segment.

[0090] In this embodiment, it is found through research that the first distance threshold and the second distance threshold are 1m and 3.5m respectively.

[0091] The camera with the smaller error in the close-range segment is regarded as the dominant camera in the close-range segment; the camera with the smaller error in the long-range segment is regarded as the dominant camera in the long-range segment;

[0092] Step 5: When actually using it, first determine the optimal distance obtained by the TOF and binocular methods;

[0093] If the optimal distances of both cameras are in the long-distance segment, the average distance value corresponding to the dominant camera in the long-distance segment is used as the final distance value.

[0094] If the optimal distances of both cameras are in the close range, the average distance of the cameras with the advantage in the close range is used as the final distance value.

[0095] If the optimal distance of any of the two distance measurements is in the middle distance segment, execute step 6;

[0096] In this embodiment, in the short distance segment of 0-1m, the binocular vision camera is dominant, and the average value of the binocular vision ranging is the final ranging value; in the long distance segment of more than 3.5m, the TOF camera is dominant, and the average value of the TOF camera ranging is the final ranging value;

[0097] Step 6: In the intermediate distance segment, the ranging results of binocular vision and TOF camera are unstable. Both may affect the ranging accuracy at a certain position due to system uncertainty, environmental interference and invalid data. Therefore, different weights are assigned to the average value of TOF ranging and the average value of binocular vision ranging. The average value of TOF camera and binocular vision ranging is multiplied by their respective weights and then summed, as shown in the following example: Figure 3 As shown, it is the final distance measurement value of the intermediate distance segment.

[0098] Taking a certain measurement location as an example, the specific steps are as follows:

[0099] Step 6.1: The TOF camera is denoted as p, and the data after the Gaussian model is filtered is X p There are r of them, with a mean of The binocular vision camera is recorded as q, and the data after the Gaussian model is filtered is X q There are r of them, with a mean of

[0100] Step 6.2: Generally speaking, these measurements consist of the true signal X and the observation error V p and V q Composition, denoted as X p =X+V p and X q =X+V q Observation error V p and V q It can be regarded as zero-mean stationary noise. Based on this, at a certain measurement position, the data measured by camera p has a variance of σ 2 =E(V p 2 ), where E(·) is the expectation. Since the TOF and binocular vision cameras are independent of each other and the ranging process does not affect each other, the observation errors between the two cameras are uncorrelated, the mean of the observation errors is 0 and is uncorrelated with the true signal. The cross-correlation function R between any cameras p and q is pq , and the autocorrelation function R of camera p pp The formula is as follows:

[0101] R pq =E(X p X q )=E(X 2 )

[0102] R pp =E(X p X p )=E(X 2 )+E(V p 2 )

[0103] Step 6.3: Since both cameras have r data after Gaussian model processing, the autocorrelation and cross-correlation function calculation formulas are as follows:

[0104]

[0105]

[0106] Where r is the number of sampling points, that is, the number of data;

[0107] Step 6.4: Variance of camera p The variances of q and camera q are:

[0108]

[0109]

[0110] Step 6.5: According to the extreme value theory of multivariate function, the weighting factor W corresponding to p and q when the total root mean square error is minimized is p and W q They are:

[0111]

[0112]

[0113] Step 6.5: The calculation formula of the fused estimated value X is as follows:

[0114]

[0115] in, and It is the mean value of the data of TOF camera and binocular vision after being filtered by Gaussian model.

[0116] In summary, the present invention provides a ranging method based on the fusion of TOF camera and binocular vision data. The TOF camera and binocular vision ranging results are fused through two fusion methods: Gaussian model processing and adaptive weighting. This suppresses the influence of errors, expands the camera's usage conditions, and improves reliability and accuracy. Specific implementation method 2:

[0118] This embodiment is a computer storage medium, which stores at least one instruction. The at least one instruction is loaded and executed by a processor to implement the distance measurement method based on the fusion of TOF camera and binocular vision data.

[0119] It should be understood that the instructions include computer program products, software, or computerized methods corresponding to any method described in the present invention; the instructions can be used to program a computer system or other electronic device. Computer storage media may include readable media on which instructions are stored, and may include but are not limited to magnetic storage media, optical storage media; magneto-optical storage media include read-only memory ROM, random access memory RAM, erasable programmable memory (e.g., EPROM and EEPROM) and flash memory layers, or other types of media suitable for storing electronic instructions. Specific implementation method three:

[0121] This embodiment is a ranging device based on the fusion of TOF camera and binocular vision data. The device includes a processor and a memory. It should be understood that, including any device including a processor and a memory described in the present invention, the device may also include other units and modules that perform display, interaction, processing, control, and other functions through signals or instructions;

[0122] At least one instruction is stored in the memory, and the at least one instruction is loaded and executed by the processor to implement the distance measurement method based on the fusion of TOF camera and binocular vision data.

[0123] The above examples are merely illustrative of the calculation model and process of the present invention and are not intended to limit the embodiments of the present invention. Persons skilled in the art will readily appreciate that other variations or modifications based on the above description are possible. This list of embodiments is not exhaustive; however, any obvious variations or modifications derived from the technical solution of the present invention remain within the scope of protection of the present invention.

Claims

1. A distance measurement method based on the fusion of TOF camera and binocular vision data, characterized in that: The following steps are involved: S1. Obtain the distance data of the target to be measured by the TOF camera and the binocular vision ranging platform. The obtained data corresponds to n measurements, that is, the TOF camera and the binocular vision camera each obtain n distance values ​​D i , i=1,2,…,n; The TOF camera and binocular vision ranging platform includes a TOF camera and a binocular vision camera. The binocular vision camera includes a binocular vision right camera and a binocular vision left camera. The centers of the TOF camera and the left and right binocular vision cameras are in a straight line, and this straight line and the line connecting the platform and the target are perpendicular to each other. S2. Perform the following processing for TOF camera and binocular vision respectively: S2.

1. Determine the Gaussian distribution function of the random measurement value d: Among them, σ is the standard deviation, d is the n distance value D i The random measurement value in, μ is the mathematical expectation, σ 2 is n distance values ​​D i The corresponding variance; S2.2: Determine the range of selectable values ​​and the critical value of the Gaussian distribution function: Where u is the lower critical value of the Gaussian distribution function; When the value of the Gaussian distribution function is greater than u, the measured value is considered to have a high probability of occurrence; m is the n distance value D i The corresponding average value; S2.3: Select the distance value X retained after Gaussian model screening from the initial distance value i , the number is r; get the optimal value of ranging: where X i is the i-th value that meets the requirements, i = 1, 2, ..., r, r is the number of values ​​that meet the requirements; S3, determining whether any of the optimal distances obtained by the TOF and binocular methods is between the first distance threshold and the second distance threshold; if so, executing S4; S4. The average value of TOF distance measurement and the average value of binocular vision distance measurement are fused according to the weights to obtain the final distance measurement value: S4.

1. Let the TOF camera be denoted as p, and the data after the Gaussian model is denoted as X p , the mean is The binocular vision camera is recorded as q, and the data after the Gaussian model is recorded as X q , the mean is S4.

2. Assume that the actual distance is X ture , the observation error between the average value of TOF ranging and binocular vision ranging is recorded as V p and V q , then X p =X ture +V p and X q =X ture +V q Observation error V p and V q Considered as zero-mean stationary noise; The data measured by camera p has a variance of σ 2 =E(V p 2 ), where E(·) is the expectation; S4.

3. Since both cameras have r data after being processed by the Gaussian model, the cross-correlation function R between camera p and camera q is obtained. pq , and the autocorrelation function R of camera p pp : S4.

4. Determine the variance of camera p And the variance of camera q: S4.

5. Determine the weighting factor W of camera p and camera q p and W q : S4.

5. Using the weighting factor W p and W q After fusion, the estimated value of the fused distance X is as follows: in, and It is the mean value of the data of TOF camera and binocular vision after being filtered by Gaussian model.

2. A distance measurement method based on TOF camera and binocular vision data fusion according to claim 1, characterized in that, The process of determining the first distance threshold and the second distance threshold is as follows: A1. Obtain the distance data of the target to be measured by the TOF camera and the binocular vision ranging platform. The obtained data corresponds to L / k measurement positions, and each position corresponds to n measurement data; L / k measurement positions: Within the range L of the target, measurement positions separated by a distance of k are placed close to the target and the distance to the target is measured in sequence. Each measurement position is measured n times, for a total of L / k measurement positions. A2. Process the TOF camera and binocular vision separately, and use the Gaussian model to process each set of data to obtain the optimal distance measurement value; A3. Based on L / k measurement positions, determine the optimal distance and the corresponding distance error, and obtain the error-distance curves of the TOF and binocular vision methods within the ranging range of 0 to L, respectively; plot the error-distance curves of the TOF and binocular vision methods together, and the intersection of the minimum distance value and the maximum distance value of the two curves is the first distance threshold and the second distance threshold.

3. A distance measurement method based on TOF camera and binocular vision data fusion according to claim 2, characterized in that, The processing of step A2 is the same as that of step S2.

4. A distance measurement method based on TOF camera and binocular vision data fusion according to claim 3, characterized in that: The first distance threshold and the second distance threshold are 1 m and 3.5 m respectively.

5. The distance measurement method based on the fusion of TOF camera and binocular vision data according to claim 2, 3 or 4, characterized in that: The method further comprises the following steps: Based on the target location, the distance within the first distance threshold is regarded as the short distance segment, and the distance outside the second distance threshold is regarded as the long distance segment; In the process of determining in S3 whether any of the optimal distances of the ranging values ​​obtained by the TOF and binocular methods is between the first distance threshold and the second distance threshold, if the optimal distances of the ranging values ​​obtained by the TOF and binocular methods are both in the long-distance segment, the average distance value corresponding to the camera with the smaller distance error between the TOF camera and the binocular vision camera in the long-distance segment is selected as the final ranging value; if the optimal distances of the ranging values ​​obtained by the TOF and binocular methods are both in the short-distance segment, the average distance value corresponding to the camera with the smaller distance error between the TOF camera and the binocular vision camera in the short-distance segment is selected as the final ranging value.

6. The distance measurement method based on the fusion of TOF camera and binocular vision data according to claim 5, characterized in that: The distance error of the camera with the smaller distance error in the long-distance segment is determined in the process of determining the ranging optimal value distance and the corresponding distance error in A3.

7. The distance measurement method based on the fusion of TOF camera and binocular vision data according to claim 6, characterized in that: The distance error of the camera with the smaller distance error in the close-range segment is determined in the process of determining the ranging optimal value distance and the corresponding distance error in A3.

8. The distance measurement method based on the fusion of TOF camera and binocular vision data according to claim 7, characterized in that: In the close-range segment, the average distance measurement value corresponding to the binocular vision camera is selected as the final distance measurement value; in the long-range segment, the average distance measurement value corresponding to the TOF camera is selected as the final distance measurement value.

9. A computer storage medium, characterized in that The storage medium stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement a ranging method based on the fusion of TOF camera and binocular vision data as described in any one of claims 1 to 8.

10. A distance measurement device based on the fusion of TOF camera and binocular vision data, characterized in that: The device includes a processor and a memory, wherein the memory stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement a ranging method based on the fusion of a TOF camera and binocular vision data as described in any one of claims 1 to 8.

Images