Image Recognition Method, System, Image Processing Device and Readable Storage Medium Applied to Self-Propelled Equipment

Abstract

The present invention discloses an image recognition method, system, image processing device and readable storage medium applied to a self-propelled device; the image recognition method applied to the self-propelled device includes the following steps: obtaining an image of the environment in the traveling direction of the self-propelled device; performing partition processing on the image according to a preset partition rule to obtain a plurality of target blocks; counting the partition feature values of the target blocks; comparing the partition feature values with a preset threshold value to obtain a recognition result. The present invention partitions the image, performs comparison processing through the partition feature values and the preset threshold value, and the recognition result obtained by using the partition feature values can effectively identify abnormalities in the target blocks, thereby reducing the misjudgment of the image.

Description

Technical Field

[0001] The present invention relates to an image recognition method, system, image processing device and readable storage medium applied to self-propelled devices, and particularly to an image recognition method, system, image processing device and readable storage medium applied to self-propelled devices that can reduce misjudgment. Background Art

[0002] When an existing lawn mower performs image recognition, it calculates eigenvalue (the eigenvalue may include an average brightness value or an average roughness value) through the overall image, and then compares the eigenvalue with a preset threshold to identify whether the image includes a non-lawn situation. However, when recognizing through the overall image, the recognition result of the eigenvalue cannot reflect the presence of obstacles, resulting in misjudgment of the image. For example, the image includes a lawn, as well as obstacles or boundaries. However, the proportion of obstacles or boundaries in the image is small. After the eigenvalue obtained by calculating through the overall image is compared and processed, the recognition result is that the image does not include a non-lawn situation. Summary of the Invention

[0003] The present invention provides an image recognition method, system, image processing device and readable storage medium applied to self-propelled devices that can reduce misjudgment.

[0004] The present invention provides an image recognition method applied to self-propelled devices; the method includes the following steps:

[0005] Obtain an image of the environment of the self-propelled device in its traveling direction;

[0006] Perform zoning processing on the image according to a preset zoning rule to obtain a plurality of target blocks;

[0007] Statistical zoning eigenvalue of the target block;

[0008] Compare the zoning eigenvalue with a preset threshold for processing to obtain a recognition result.

[0009] Optionally, the image recognition method applied to self-propelled devices further includes: dividing the image into a plurality of preprocessing blocks, and obtaining a preprocessing eigenvalue of each preprocessing block.

[0010] Optionally, the preprocessing eigenvalue includes the number of pixels in a predetermined chromaticity interval corresponding to the preprocessing block, an average brightness value, and an average roughness value, or the preprocessing eigenvalue includes an average brightness value and an average roughness value.

[0011] Optionally, the target block includes a preprocessing block or a combination of multiple preprocessing blocks. The step of calculating the partition feature value of the target block includes: obtaining the preprocessing feature values of the preprocessing blocks in the target block, and then averaging the preprocessing feature values of the preprocessing blocks in the target block to obtain the partition feature value of the target block.

[0012] Optionally, the step of partitioning the image according to a preset partitioning rule to obtain multiple target blocks includes: partitioning the image according to multiple preset partitioning rules to obtain multiple target blocks, and the target blocks corresponding to the multiple preset partitioning rules are not completely the same.

[0013] Optionally, the image recognition method applied to the self - propelled device further includes obtaining the total number of pixels in a predetermined chromaticity interval corresponding to the image; determining whether the total number of pixels in the predetermined chromaticity interval meets a preset total pixel threshold. When the total number of pixels in the predetermined chromaticity interval meets the preset total pixel threshold, comparing the partition feature value with a first preset threshold to obtain a recognition result; when the total number of pixels in the predetermined chromaticity interval does not meet the preset total pixel threshold, obtaining a recognition result that the image is an image to be screened.

[0014] Optionally, the step of comparing the partition feature value with a preset threshold to obtain a recognition result is to compare the partition feature value with a second preset threshold to obtain a recognition result; the numerical range of the second preset threshold is larger than the numerical range of the first preset threshold.

[0015] The present invention also provides an image recognition system applied to a self - propelled device. The system includes:

[0016] An image acquisition module, which is used to acquire an image of the environment in the traveling direction of the self - propelled device;

[0017] A partitioning module, which is used to partition the image according to a preset partitioning rule to obtain multiple target blocks;

[0018] A feature processing module, which is used to calculate the partition feature value of the target block;

[0019] An image recognition module, which is used to compare the partition feature value with a preset threshold to obtain a recognition result.

[0020] The present invention also provides an image processing device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the image recognition method applied to the self - propelled device are implemented.

[0021] The present invention also provides a readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the image recognition method applied to the self-propelled device are implemented.

[0022] Compared with the prior art, the present invention partitions an image, performs comparison processing on the partition eigenvalue and a preset threshold value, and the recognition result obtained by using the partition eigenvalue can effectively identify abnormalities in the target block, thereby reducing misjudgment of the image. The present invention divides the image into preprocessing blocks, and the preprocessing eigenvalue of the preprocessing block can be used as intermediate data for calculating the partition eigenvalue. The preprocessing eigenvalue is commonly used for calculating the partition eigenvalues of different preset partitioning rules, and the partition eigenvalue can be obtained through the combination of the preprocessing features, which is beneficial to simplifying the operation of the partition eigenvalue through multiple preset partitioning rules. The present invention sets multiple preset partitioning rules to partition the image, and the target blocks corresponding to the multiple preset partitioning rules are not completely the same, so as to reduce misjudgment caused by a single preset partitioning rule. The present invention simultaneously sets multiple determination bases, and the preset threshold values corresponding to different determination bases are different, effectively reducing misjudgment caused by a single preset threshold value or a single determination basis. Description of the Drawings

[0023] Figure 1 is a flowchart of the image recognition method applied to the self-propelled device of the present invention;

[0024] Figure 2 is Figure 1 a flowchart of step S10;

[0025] Figure 3 is Figure 1 a flowchart of step S30;

[0026] Figure 4 is a schematic diagram of the image being divided into preprocessing blocks;

[0027] Figure 5 is a schematic diagram of the image being divided into target blocks according to the first preset partitioning rule;

[0028] Figure 6 is a schematic diagram of the image being divided into target blocks according to the second preset partitioning rule;

[0029] Figure 7 is a principle block diagram of the image recognition system applied to the self-propelled device of the present invention. Detailed Embodiments

[0030] To enable those skilled in the art to better understand the technical solutions in the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0031] Please refer to Figure 1-7 As shown, the present invention provides an image recognition method applied to a self-propelled device; the method includes the following steps:

[0032] Step S10: Obtain an image of the environment in the traveling direction of the self-propelled device;

[0033] Step S20: Partition the image according to a preset partitioning rule to obtain a plurality of target blocks;

[0034] Step S30: Statistically calculate the partitioning feature values of the target blocks;

[0035] Step S40: Compare the partitioning feature values with a preset threshold for processing to obtain an identification result.

[0036] In another embodiment of the present invention, the self-propelled device in step S10 is a lawn mower or a sweeper, etc., and the image in step S10 is an image obtained by the lawn mower or the sweeper in its traveling direction. When the self-propelled device is a lawn mower, the preset threshold is a threshold related to the lawn, and the identification result includes that there is a non-lawn area in the image, the image is an image to be screened, the image to be screened may be a lawn image, and the non-lawn area may be an abnormal area such as an obstacle or a boundary. Through the image recognition method applied to the self-propelled device of the present invention, the images that may be lawns can be screened out for subsequent image processing. When the self-propelled device is a sweeper, the preset threshold is a threshold related to the road surface, and the identification result includes that there is a non-road surface area in the image, the image is an image to be screened, the image to be screened may be a road surface image, and the non-road surface area may be an abnormal area such as an obstacle or a boundary.

[0037] Step S10 includes the following steps:

[0038] Step S110: Obtain the original image orgMat;

[0039] Step S120: Separate the original image orgMat to obtain an h-channel image and a v-channel image;

[0040] Step S130: Preprocess the v-channel image to obtain a preprocessed image normVMat, where the preprocessing includes filtering, normalization, etc.;

[0041] Step S140: Perform edge extraction on the preprocessed image normVMat to obtain an edge image cannyMat, where the edge extraction can use a canny operator, etc.

[0042] In another embodiment of the present invention, the image recognition method applied to the self-propelled device further includes: dividing the image 1 into several preprocessing blocks 11, and obtaining the preprocessing feature values of each preprocessing block. The preprocessing block 11 is the smallest unit module for combining into a target block, and the image can be evenly divided into multiple preprocessing blocks 11 with equal areas. For example, the image is divided into preprocessing blocks 11 of 10*10.

[0043] In another embodiment of the present invention, the preprocessing feature values include the number of pixels in a predetermined chromaticity interval corresponding to the preprocessing block, the average brightness value, and the average roughness value, or the preprocessing feature values include the average brightness value and the average roughness value.

[0044] The predetermined chromaticity interval can be set as needed. For example, in order to identify whether there is a non-lawn area in the image, the predetermined chromaticity interval can be set to 15-105.

[0045] Calculate the number of pixels in the predetermined chromaticity interval of the preprocessing block according to the h-channel image, calculate the average brightness value of the preprocessing block according to the v-channel image, and calculate the average roughness value of the preprocessing block according to the edge image. The 10*10 preprocessing block includes the preprocessing feature values corresponding to the preprocessing block. The preprocessing feature values include the number of pixels in the predetermined chromaticity interval, the average brightness value, and the average roughness value, or the preprocessing feature values include the average brightness value and the average roughness value. The preprocessing feature values are used as intermediate data to facilitate step S30 to statistically calculate the partition feature values of the target block through the preprocessing feature values.

[0046] In another embodiment of the present invention, the target block in step S20 includes a preprocessing block or a combination of multiple preprocessing blocks.

[0047] In another embodiment of the present invention, step S20 includes: partitioning the image according to multiple preset partitioning rules to obtain multiple target blocks, and the target blocks corresponding to the multiple preset partitioning rules are not completely the same. The preset partitioning rules can be non-overlapping parallel partitioning rules or partially overlapping overlapping partitioning rules, and all the target blocks of each preset partitioning rule cover and form the image.

[0048] Assume that each target block includes 2 preprocessing blocks, which are partitioned by non-overlapping parallel partitioning rules. The preprocessing blocks of the image can be combined horizontally or vertically. Then, the image 1 (including 10*10 preprocessing blocks 11) obtains a total of 50 target blocks, and each target block includes corresponding preprocessing blocks respectively.

[0049] Multiple preset partitioning rules are combined to obtain target blocks, so as to reduce misjudgment caused by a single preset partitioning rule. For example, after the image 1 is partitioned by the first preset partitioning rule, it is scattered in multiple first target blocks 12. Comparing the partitioning characteristic values of the first target blocks 12 obtained by the first preset partitioning rule with the preset threshold cannot identify that there is a non-lawn area in the image, while after the image is partitioned by the second preset partitioning rule, it is concentrated in some second target blocks 13. Comparing the partitioning characteristic values of the second target blocks 13 obtained by the second preset partitioning rule with the preset threshold can identify that there is a non-lawn area in the image 1. Multiple preset partitioning rules can be set as needed. Different preset partitioning rules determine the shape of the partition or the number of preprocessing blocks included. Therefore, the more preset partitioning rules, the more likely the misjudgment can be reduced.

[0050] In another embodiment of the present invention, the step S30 includes:

[0051] Step S310: Obtain the preprocessing characteristic values of each preprocessing block in the target block,

[0052] Step S320: Perform an averaging process on the preprocessing characteristic values of each preprocessing block in the target block to obtain the partitioning characteristic value of the target block.

[0053] After partitioning according to the first preset partitioning rule, the partitioning characteristic value of the first target block 12 includes the number of pixels GSn (n = 1,...50) in the predetermined chromaticity interval corresponding to the first target block 12, the average brightness value GBn (n = 1,...50), or the average roughness value GHn (n = 1,...50).

[0054] After partitioning according to the second preset partitioning rule, the partitioning characteristic value of the second target block 13 includes the number of pixels RSn (n = 1,...50) in the predetermined chromaticity interval corresponding to the second target block 13, the average brightness value RBn (n = 1,...50), or the average roughness value RHn (n = 1,...50).

[0055] To reduce misjudgment caused by a single preset threshold or a single determination basis, the image recognition method of the present invention applied to a self-propelled device can set multiple preset thresholds or multiple determination bases, and the combination of determination elements can be selected as needed.

[0056] In another embodiment of the present invention, identification is performed through a first preset threshold and a first determination basis, wherein the determination elements in the first determination basis include the total number of pixels and the partition characteristic value, and the partition characteristic value includes the average brightness value and the average roughness value of the target block. The image recognition method applied to the self-propelled device further includes obtaining the total number of pixels in a predetermined chromaticity interval corresponding to the image; determining whether the total number of pixels in the predetermined chromaticity interval meets a preset total pixel threshold, and when the total number of pixels in the predetermined chromaticity interval meets the preset total pixel threshold, comparing the partition characteristic value with the first preset threshold to obtain an identification result; when the total number of pixels in the predetermined chromaticity interval does not meet the preset total pixel threshold, obtaining an identification result that the image is an image to be screened.

[0057] The total number of pixels S in the predetermined chromaticity interval corresponding to the image can be obtained by summing the number of pixels Sn in the preprocessing blocks, that is, the total number of pixels S in the predetermined chromaticity interval = S1 + … + S100, where Sn (n = 1,...100) is the number of pixels in the preprocessing block in the predetermined chromaticity interval, or can also be directly counted from the image before partitioning.

[0058] In another embodiment of the present invention, identification is performed through a second preset threshold and a second determination basis, wherein the determination elements in the second determination basis include the partition characteristic value but do not include the total number of pixels, and the partition characteristic value includes the number of pixels of the target block, the average brightness value and the average roughness value of the target block. The comparing the partition characteristic value with a preset threshold to obtain an identification result is to compare the partition characteristic value with the second preset threshold to obtain an identification result; the numerical range of the second preset threshold is larger than the numerical range of the first preset threshold.

[0059] Taking the total number of pixels in the predetermined chromaticity interval as one of the determination elements, when combined with the total number of pixels in the predetermined chromaticity interval for determination, the numerical range of the first preset threshold used for comparison processing with the average brightness value GBn (RBn) and the average roughness value GHn of the target block is smaller.

[0060] Taking the number of pixels GSn (RSn) of the target block as one of the determination elements, when combined with the number of pixels of the target block for determination, the numerical range of the second preset threshold used for comparison processing with the average brightness value GBn and the average roughness value GHn of the target block is larger.

[0061] Determination basis 1: S > 18000 and 36 < GBn (RBn) <= 110 and GHn (RHn) < 0.025

[0062] Determination criterion 2: GHn(RHn) < 0.09 and GBn(RBn) > 36 and GSn(RSn) < 4100

[0063] In the process of identifying whether there is a non-lawn area in the image, if the image meets Determination criterion 1 or the image meets Determination criterion 2, then there is a non-lawn area in the image; if the image does not meet Determination criterion 1 and does not meet Determination criterion 2, then the image is an image to be screened, and the image to be screened may be a lawn image, and subsequent image processing (judgment of other feature values in the image to be screened) is required to further identify whether there is a non-lawn area in the image.

[0064] The present invention also provides an image recognition system 300 applied to a self-propelled device, and the system 300 includes:

[0065] An image acquisition module 31, which is used to acquire an image of the environment in the traveling direction of the self-propelled device;

[0066] A partitioning module 32, which is used to partition the image according to a preset partitioning rule to obtain a plurality of target blocks;

[0067] A feature processing module 33, which is used to count the partitioning feature values of the target blocks;

[0068] An image recognition module 34, which is used to compare the partitioning feature values with preset thresholds to obtain a recognition result.

[0069] The present invention also provides an image processing device, including a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps of the image recognition method applied to the self-propelled device are implemented.

[0070] The present invention also provides a readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the image recognition method applied to the self-propelled device are implemented.

[0071] In summary, the present invention partitions an image, compares and processes the partition eigenvalue with a preset threshold, and the recognition result obtained by using the partition eigenvalue can effectively identify the abnormality in the target block, thereby reducing the misjudgment of the image. The present invention divides the image into preprocessing blocks, and the preprocessing eigenvalue of the preprocessing block can be used as intermediate data for calculating the partition eigenvalue. The preprocessing eigenvalue is commonly used for calculating the partition eigenvalues of different preset partitioning rules. By combining the preprocessing features, the partition eigenvalue can be obtained, which is beneficial to simplifying the operation of the partition eigenvalue through multiple preset partitioning rules. The present invention sets multiple preset partitioning rules to partition the image, and the target blocks corresponding to the multiple preset partitioning rules are not completely the same, so as to reduce the misjudgment caused by a single preset partitioning rule. The present invention also sets multiple determination bases, and the preset thresholds corresponding to different determination bases are different, effectively reducing the misjudgment caused by a single preset threshold or a single determination basis.

[0072] In addition, it should be understood that although this specification is described according to the embodiments, not every embodiment only contains an independent technical solution. This narrative way of the specification is only for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0073] The series of detailed descriptions listed above are only specific descriptions of the feasible embodiments of the present invention, and are not used to limit the protection scope of the present invention. Any equivalent embodiments or changes made without departing from the technical spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. An image recognition method applied to a self-propelled device; characterized in that, The method includes the following steps: Obtain an image of the environment in the traveling direction of the self-propelled device; Partition the image according to a preset partitioning rule to obtain a plurality of target blocks, and the plurality of target blocks corresponding to the preset partitioning rule are not completely the same; Statistically analyze the partitioning feature values of the target blocks; Compare the partitioning feature values with a preset threshold to obtain an identification result; The image recognition method applied to the self-propelled device further includes: dividing the image into a plurality of preprocessing blocks, and obtaining the preprocessing feature values of each preprocessing block, where the preprocessing feature values include the number of pixels in a predetermined chromaticity interval corresponding to the preprocessing block, the average brightness value, and the average roughness value, or the preprocessing feature values include the average brightness value and the average roughness value; The preprocessing feature values are obtained through the following method: Separate the image to obtain an h-channel image and a v-channel image; Preprocess the v-channel image to obtain a preprocessed image; Extract edges from the preprocessed image to obtain an edge image; Calculate the number of pixels in the predetermined chromaticity interval of the preprocessing block according to the h-channel image, calculate the average brightness value of the preprocessing block according to the v-channel image, and calculate the average roughness value of the preprocessing block according to the edge image; The target block includes one preprocessing block or a combination of multiple preprocessing blocks, and the step of statistically analyzing the partitioning feature values of the target block includes: obtaining the preprocessing feature values of each preprocessing block in the target block, and then averaging the preprocessing feature values of each preprocessing block in the target block to obtain the partitioning feature value of the target block; The image recognition method applied to the self-propelled device further includes: obtaining the total number of pixels in the predetermined chromaticity interval corresponding to the image; determining whether the total number of pixels in the predetermined chromaticity interval meets a preset total pixel threshold, and when the total number of pixels in the predetermined chromaticity interval meets the preset total pixel threshold, compare the partitioning feature values with a first preset threshold to obtain an identification result; When the total number of pixels in the predetermined chromaticity interval does not meet the preset total pixel threshold, obtain an identification result that the image is an image to be screened; The step of comparing the partitioning feature values with a preset threshold to obtain an identification result includes: Compare the partitioning feature values with a second preset threshold to obtain an identification result; wherein, the numerical range of the second preset threshold is larger than the numerical range of the first preset threshold.

2. An image recognition system applied to a self-propelled device, characterized in that, The system includes: An image acquisition module, which is used to obtain an image of the environment in the traveling direction of the self-propelled device; A partitioning module, which is used to partition the image according to a preset partitioning rule to obtain a plurality of target blocks, and the plurality of target blocks corresponding to the preset partitioning rule are not completely the same; A feature processing module, which is used to statistically analyze the partitioning feature values of the target blocks; An image recognition module, which is used to compare the partitioning feature values with a preset threshold to obtain an identification result; The feature processing module is further configured to: divide the image into a plurality of preprocessing blocks, and obtain preprocessing feature values of each of the preprocessing blocks, where the target block includes one preprocessing block or a combination of a plurality of preprocessing blocks, and the statistical partition feature value of the target block includes: obtaining the preprocessing feature values of the preprocessing blocks in the target block, and then performing an averaging process on the preprocessing feature values of the preprocessing blocks in the target block to obtain the partition feature value of the target block; Wherein, the preprocessing feature value includes the number of pixels in a predetermined chromaticity interval corresponding to the preprocessing block, an average brightness value, and an average roughness value, or the preprocessing feature value includes the average brightness value and the average roughness value; the preprocessing feature value is obtained by the following method: separating the image to obtain an h-channel image and a v-channel image; preprocessing the v-channel image to obtain a preprocessing image; performing edge extraction on the preprocessing image to obtain an edge image; calculating the number of pixels in the predetermined chromaticity interval of the preprocessing block according to the h-channel image, calculating the average brightness value of the preprocessing block according to the v-channel image, and calculating the average roughness value of the preprocessing block according to the edge image; The image recognition module is further configured to: obtain the total number of pixels in a predetermined chromaticity interval corresponding to the image; determine whether the total number of pixels in the predetermined chromaticity interval meets a preset total pixel threshold, and when the total number of pixels in the predetermined chromaticity interval meets the preset total pixel threshold, compare the partition feature value with a first preset threshold to obtain an identification result; when the total number of pixels in the predetermined chromaticity interval does not meet the preset total pixel threshold, obtain an identification result that the image is an image to be screened; compare the partition feature value with a second preset threshold to obtain an identification result; wherein, the numerical range of the second preset threshold is larger than the numerical range of the first preset threshold.

3. An image processing device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the image recognition method applied to the self-propelled device according to claim 1 are implemented.

4. A readable storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by a processor, the steps of the image recognition method applied to the self-propelled device according to claim 1 are implemented.

Images