Air conditioner and target tracking method and equipment

Abstract

The invention discloses an air conditioner and a target tracking method and device.The air conditioner is provided with an image collecting module, and current image information collected by the image collecting module is obtained in real time; identifying all targets on the current image information; when the target is identified for the first time, arranging and numbering the position information of the target on the current image information according to a preset sequence to form a target sequence; when the target is not identified for the first time, arranging the position information of the target on the current image information according to a preset sequence; and matching the arranged position information of the targets on the current image information with the position information of the target sequence, and updating the target sequence according to a matching result. According to the invention, after the target is identified, the dynamic tracking of the target is completed by matching with the historical position, the calculation is simple, the reliability is high, and the calculation burden of a common tracking algorithm is obviously reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of air conditioners, and particularly to an air conditioner, a target tracking method and device. Background Art

[0002] At present, air conditioners have become essential household appliances in people's daily lives, with various functions such as refrigeration, heating, and dehumidification. With the development of technology, intelligent air conditioners are equipped with functions to monitor the user's state. When detecting the user's state, the air conditioner includes two parts: identification and tracking, which require identifying the user and tracking the user's position to analyze the changes in the user's state.

[0003] Most of the existing target tracking technologies are based on methods of large amounts of data and deep neural networks. However, the existing technologies have at least the following problems: The tracking methods based on deep neural networks have high training costs and usage costs and cannot be well applied to the terminal processing devices carried by air conditioners. Summary of the Invention

[0004] The purpose of the embodiments of the present invention is to provide an air conditioner, a target tracking method and device, which, after identifying the target, complete the dynamic tracking of the target by matching with the historical position, with simple calculation and high reliability, and significantly reduce the calculation burden of common tracking algorithms.

[0005] To achieve the above purpose, the embodiments of the present invention provide an air conditioner, including:

[0006] A housing;

[0007] An image acquisition module, disposed on the housing, for acquiring image information of the current indoor environment;

[0008] A controller, configured to:

[0009] Acquire the current image information acquired by the image acquisition module in real time;

[0010] Identify all targets on the current image information;

[0011] When a target is identified for the first time, arrange and number the position information of the target on the current image information in a preset order to form a target sequence;

[0012] When a target is not identified for the first time, arrange the position information of the target on the current image information in a preset order;

[0013] Match the arranged position information of the target on the current image information with the position information of the target sequence, and update the target sequence according to the matching result.

[0014] As an improvement to the above solution, the preset order is the order in which the abscissa of the position information of the target increases from small to large.

[0015] As an improvement to the above solution, the matching of the position information of the target on the arranged current image information with the position information of the target sequence and the updating of the target sequence according to the matching result include:

[0016] Generate a first vector according to the position information of the target on the arranged current image information;

[0017] Generate a second vector according to the position information of the target in the current target sequence;

[0018] Construct an intersection over union matrix by calculating the intersection over union between the first vector and the second vector;

[0019] Update the number of the target corresponding to each position information in the first vector according to the intersection over union matrix to obtain an updated target sequence.

[0020] As an improvement to the above solution, the updating of the number of the target corresponding to each position information in the first vector according to the intersection over union matrix to obtain an updated target sequence includes:

[0021] Determine the same target on the current image information as the target in the target sequence according to the intersection over union matrix;

[0022] Bind the number of the same target in the target sequence to the same target on the current image information;

[0023] Add a new number to the target that is different from the target sequence on the current image information;

[0024] Until all the targets in the current image information are bound with numbers or new numbers are added, an updated target sequence is obtained.

[0025] As an improvement to the above solution, the updating of the number of the target corresponding to each position information in the first vector according to the intersection over union matrix to obtain an updated target sequence includes:

[0026] Obtain the target corresponding to the position information of the first vector corresponding to the maximum value in the intersection over union matrix, denoted as the current target, and obtain the target corresponding to the position information of the second vector corresponding to the maximum value in the intersection over union matrix, denoted as the historical target;

[0027] Determine that the current target and the historical target are the same target, and bind the number of the historical target to the current target as the number of the current target;

[0028] Set the position information of the column vector and row vector corresponding to the maximum value in the intersection over union matrix to 0, and recalculate the intersection over union matrix;

[0029] Until all the targets corresponding to the position information in the first vector are bound with numbers to obtain an updated target sequence, or until there is no intersection over union with a single maximum value or the maximum value is less than a preset intersection over union threshold, assign new numbers to the targets corresponding to the remaining position information in the first vector to obtain an updated target sequence.

[0030] As an improvement to the above solution, identifying all the targets on the current image information includes:

[0031] Input the current image information into a preset contour recognition model for recognition to obtain the contour information of all candidate targets output by the contour recognition model;

[0032] Input the current image information into a preset key feature recognition model for recognition to obtain the key feature information of all candidate targets output by the key feature recognition model;

[0033] Identify all the targets on the current image information according to the contour information and the key feature information.

[0034] As an improvement to the above solution, the contour information is the position information of the contour frame of the candidate target; the key feature information is the position information of the key feature points of the candidate target;

[0035] Then identifying all the targets on the current image information according to the contour information and the key feature information includes:

[0036] According to the position information of the contour frame and the position information of the key feature points, when it is determined that any one of the key feature points is located within any one of the contour frames, regard the candidate target corresponding to the any one contour frame as a target to identify all the targets on the current image information.

[0037] As an improvement to the above solution, identifying all the targets on the current image information further includes:

[0038] According to the position information of all the targets, when it is determined that the distance between any two targets is less than or equal to a preset distance threshold, select one of the any two targets to be retained and delete the other target to update all the targets identified on the current image information.

[0039] An embodiment of the present invention further provides a target tracking method, including:

[0040] Obtain the current image information collected by a preset image acquisition module in real time;

[0041] Identify all targets on the current image information;

[0042] When a target is recognized for the first time, arrange and number the position information of the target on the current image information in a preset order to form a target sequence;

[0043] When a target is not recognized for the first time, arrange the position information of the target on the current image information in a preset order;

[0044] Match the arranged position information of the target on the current image information with the position information of the target sequence, and update the target sequence according to the matching result.

[0045] An embodiment of the present invention further provides a target tracking device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the target tracking method as described above is implemented.

[0046] Compared with the prior art, the air conditioner, target tracking method, and device disclosed in the present invention identify and track targets by collecting image information of the current environment in real time. After identifying the targets, a method of matching with historical positions is used to complete the dynamic tracking of the targets and record the state changes of the targets. The calculation is simple and the reliability is high, significantly reducing the calculation burden of common tracking algorithms, and being able to significantly reduce the adverse effects caused by errors such as target misrecognition and under-identification of targets, improving the accuracy of the operation control of the air conditioner and the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a schematic structural diagram of an air conditioner provided by an embodiment of the present invention;

[0048] Figure 2 is a schematic structural diagram of a split air conditioner in an embodiment of the present invention;

[0049] Figure 3 is a schematic structural diagram of an integrated air conditioner in an embodiment of the present invention;

[0050] Figure 4 is a schematic structural diagram of a refrigeration system in an embodiment of the present invention;

[0051] Figure 5 is a schematic diagram of the working process of a controller in an embodiment of the present invention;

[0052] Figure 6 is a schematic diagram of the principle of the matching result of under-identification of targets in an embodiment of the present invention;

[0053] Figure 7 It is a schematic diagram of the principle of the matching result of target mis-identification in the embodiments of the present invention;

[0054] Figure 8 It is a schematic diagram of the principle of the matching result of target increase or decrease in the embodiments of the present invention;

[0055] Figure 9 It is a schematic diagram of the principle of identifying a target in the embodiments of the present invention.

[0056] Figure 10 It is a schematic flow diagram of a target tracking method provided by the embodiments of the present invention;

[0057] Figure 11 It is a schematic structural diagram of a target tracking device provided by the embodiments of the present invention. Detailed implementation manners

[0058] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described 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. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0059] In the description of the present application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation to the present application.

[0060] The terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present application, unless otherwise stated, the meaning of "plurality" is two or more.

[0061] In the description of the present application, it should be noted that unless otherwise clearly specified and defined, the terms "installation", "connection", and "coupling" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0062] See Figure 1 , which is a schematic structural diagram of an air conditioner provided by an embodiment of the present invention. The embodiment of the present invention provides an air conditioner 10, which can be used to perform functions such as refrigeration, heating, and dehumidification. In an alternative embodiment, the air conditioner is a split-type air conditioner. See Figure 2 , which is a schematic structural diagram of a split-type air conditioner in an embodiment of the present invention. The split-type air conditioner includes an indoor unit 101 and an outdoor unit 102. The indoor unit includes a blowing device such as an air outlet. The outdoor unit includes a refrigeration system composed of devices such as a compressor, an evaporator, and a condenser. The specific structures and working principles of the indoor and outdoor units of the split-type air conditioner can refer to the prior art and will not be elaborated here. In another embodiment, the air conditioner is an integrated air conditioner. See Figure 3 , which is a schematic structural diagram of an integrated air conditioner in an embodiment of the present invention. The integrated air conditioner includes an indoor unit 103. The indoor unit integrates a blowing device such as an air outlet and a functional system composed of a compressor, etc. The specific structure and working principle of the integrated air conditioner can refer to the prior art and will not be elaborated here.

[0063] Specifically, the air conditioner includes a refrigeration system 11. See Figure 4 , which is a schematic structural diagram of the refrigeration system in an embodiment of the present invention. The refrigeration system includes components such as a compressor 111, an outdoor heat exchanger 112, a throttling component 113, and an indoor heat exchanger 114. Through the refrigeration system, the refrigerant is transported in different directions, thereby entering the refrigeration mode or the heating mode to achieve the refrigeration function or the heating function.

[0064] The air conditioner further includes an image acquisition module 12, which is arranged on the housing. The image acquisition module 12 is used to acquire image information of the current indoor environment.

[0065] Optionally, the image acquisition module 12 can be a camera, an infrared light sensor, etc.

[0066] The air conditioner further includes a controller 13. See Figure 5 , which is a schematic working process diagram of the controller in an embodiment of the present invention. The controller 13 is used to execute steps S11 to S15:

[0067] S11. Obtain the current image information collected by the image acquisition module in real time;

[0068] S12. Identify all the targets on the current image information;

[0069] S13. When a target is identified for the first time, arrange and number the position information of the target on the current image information in a preset order to form a target sequence;

[0070] S14. When a target is not identified for the first time, arrange the position information of the target on the current image information in a preset order;

[0071] S15. Match the arranged position information of the target on the current image information with the position information of the target sequence, and update the target sequence according to the matching result.

[0072] In the embodiment of the present invention, the controller 13 obtains the current image information collected by the image acquisition module 12 in real time, and performs target recognition work according to the current image information, that is, identifies all the targets on the current image information.

[0073] Optionally, the target is a human body.

[0074] It should be noted that the target recognition work can adopt the target recognition means in the prior art. For example, the user can be recognized by using a preset recognition model to recognize the human body contour or human body parts, which is not specifically limited herein.

[0075] Further, the controller performs target tracking work according to the recognized target. The following is explained in two cases:

[0076] In one case, when a target is recognized for the first time, that is, there is no historical target sequence information, a historical list is created, the position information of all the targets recognized on the current image information is arranged and saved in a preset order, and all the targets are numbered in order, and the order information of the targets, that is, the target numbers, is output to form a target sequence.

[0077] In another case, when a target is recognized again, the position information of all the targets recognized on the current image information is arranged in a preset order and matched with the target sequence of the historical list, so that the original targets are matched and numbered and bound, and the position information and new numbers are added to the newly appeared targets. The target sequence of the historical list is updated according to the matching result and output.

[0078] Based on this, according to the target sequence, the state of the target is saved in the corresponding scalar to achieve real-time tracking and change perception of the target state.

[0079] By adopting the technical means of the embodiment of the present invention, the target is recognized and tracked by real-time collecting the image information of the current environment. After the target is recognized, the dynamic tracking of the target and the recording of the state change of the target are completed by the method of matching with the historical position. The calculation is simple and the reliability is high, significantly reducing the calculation burden of the common tracking algorithm, and being able to significantly reduce the adverse effects brought by errors such as target misrecognition and under-recognition of the target, improving the accuracy of the operation control of the air conditioner and the user experience.

[0080] As a preferred embodiment, the embodiment of the present invention is further implemented on the basis of the above embodiment, and the preset order is the order of the abscissa of the position information of the target from small to large.

[0081] Specifically, at the beginning of recognition, there may be a situation where the target cannot be recognized or there is no target in the image. At this time, this picture is skipped until the target is recognized. When the user appears for the first time, the position information of all targets is recognized and stored in the historical user variable pl, and sorted and numbered o in the left-right order of the abscissa. Sorting and numbering can assist the subsequent matching method to ensure that the user order will not be wrong due to the recognition result and improve the accuracy of tracking.

[0082] When the user is recognized again, the position information of all users is sorted in the left-right order of the abscissa and stored in the new user variable pn. Then, the number o of the new user is determined by the matching method of the new user variable pn and the historical user variable pl, so as to store the information of the new user in the correct position.

[0083] As a preferred embodiment, step S15, that is, matching the position information of the target on the arranged current image information with the position information of the target sequence and updating the target sequence according to the matching result, includes:

[0084] S151. Generate a first vector according to the position information of the target on the arranged current image information;

[0085] S152. Generate a second vector according to the position information of the target in the current target sequence;

[0086] S153. Construct an intersection-over-union matrix by calculating the intersection-over-union between the first vector and the second vector;

[0087] S154. Update the number of the target corresponding to each position information in the first vector according to the intersection-over-union matrix to obtain an updated target sequence.

[0088] To accurately achieve this goal, the embodiments of the present invention are designed as follows: First, an intersection over union matrix is set up to calculate and store the intersection over union between the current new target variable pn and the historical target variable pl, and then the target number o is updated according to the result of the intersection over union to obtain an updated target sequence.

[0089] As a preferred embodiment, step S154, that is, updating the numbers of the targets corresponding to each position information in the first vector according to the intersection over union matrix to obtain an updated target sequence, includes:

[0090] Determine the same target on the current image information as the target in the target sequence according to the intersection over union matrix;

[0091] Bind the numbers of the same target on the current image information to the same target in the target sequence;

[0092] For the targets on the current image information that are all different from the target sequence, new numbers are added;

[0093] Until all the targets in the current image information are bound with numbers or new numbers are added to obtain an updated target sequence.

[0094] In the embodiments of the present invention, by calculating the intersection over union matrix, the distance between the targets on the current image information and the targets in the historical target sequence is determined, so as to determine whether they are the same target. For the same target, the relevant number bindings are updated, and for the newly added targets, new numbers are added to update the target sequence.

[0095] More preferably, step S154, that is, updating the numbers of the targets corresponding to each position information in the first vector according to the intersection over union matrix to obtain an updated target sequence, includes:

[0096] Obtain the target corresponding to the position information of the first vector corresponding to the maximum value in the intersection over union matrix, denoted as the current target, and obtain the target corresponding to the position information of the second vector corresponding to the maximum value in the intersection over union matrix, denoted as the historical target;

[0097] Determine the current target and the historical target as the same target, and bind the number of the historical target to the current target as the number of the current target;

[0098] Set the position information of the column vector and row vector corresponding to the maximum value in the intersection over union matrix to 0, and recalculate the intersection over union matrix;

[0099] Until all the targets corresponding to the position information in the first vector are bound with numbers to obtain an updated target sequence, or until there is no intersection over union with a single maximum value or the maximum value is less than a preset intersection over union threshold, new numbers are added to the targets corresponding to the remaining position information in the first vector to obtain an updated target sequence.

[0100] In the embodiments of the present invention, assume that the pl of the historical target variable has 4 sets of values such as l1, l2, l3, l4, and the target numbers are [1, 2, 3, 4]; the pn of the current target variable has 3 sets of values such as n1, n2, n3, then the intersection over union matrix is 4×3, where the rows represent the historical data and the columns represent the current data. The intersection over union between the two variables pl and pn is calculated in turn and saved in the intersection over union matrix. For example, the intersection over union between l1 and n1 is saved in the first row and the first column, and the intersection over union between l1 and n2 is saved in the first row and the second column, and so on. Among them, the larger the intersection over union, the closer the distance between the two positions.

[0101] Further, take the maximum value in the intersection over union matrix. The row and column corresponding to this maximum value represent the closest historical variable and current variable. For example, the value in the second row and the third column is the largest, indicating that the distance between l2 and n3 is the largest. Therefore, the current variable n3 is bound with the number 2 of the historical variable l2, that is, in the updated target sequence, the third value of the target number o is 2. (The number of values in the target number is the same as the number of current target variables)

[0102] Further, set all the values in the row and column corresponding to the maximum value to 0, that is, the values in the second row and the third column no longer participate in the comparison. Repeat the step of calculating the intersection over union matrix, and take the maximum value in the intersection over union matrix. For example, the value in the first row and the first column is the largest, indicating that the distance between l1 and n1 is the largest. Therefore, the current variable n1 is bound with the number 1 of the historical variable l1, that is, in the updated target sequence, the first value of the target number o is 1.

[0103] Further, set all the values in the row and column corresponding to the maximum value to 0, that is, the values in the first row and the first column no longer participate in the comparison. Repeat the step of calculating the intersection over union matrix. If there is no longer a single maximum value in the intersection over union matrix at this time, or the maximum value is less than the preset threshold, it can be considered that n2 in the new target variable is a newly added target, then a new number is added and determined to be number 5. Therefore, the target number of the updated target sequence is [1, 5, 3]. Furthermore, the historical target variable pl is updated according to the user number o and the current target variable pn.

[0104] See Figures 6 to 8 , Figure 6 is a schematic diagram of the principle of the matching result of target under-identification in the embodiments of the present invention, Figure 7 is a schematic diagram of the principle of the matching result of target mis-identification in the embodiments of the present invention,Figure 8 This is a schematic diagram of the principle of the matching result of the target increase and decrease in the embodiment of the present invention. It can be seen that the use of the target number o can effectively avoid errors caused by recognition accuracy problems. When the recognition result is not completely correct, situations such as under-recognition or mis-recognition may occur, or the increase and decrease of users in the sensor may be caused by the movement of the target. For example, if there are two targets in the image and both are recognized, the targets can be sorted and numbered in sequence at this time, and the required information can be stored in the corresponding positions; if one of the targets may not be recognized or mis-recognized due to recognition errors, the embodiment of the present invention can also ensure that the recognized targets are correctly numbered to avoid tracking errors. In addition, when the target increases, the embodiment of the present invention can actively expand the target sequence according to the matching result.

[0105] As a preferred embodiment, the embodiment of the present invention further optimizes the target recognition work. Then, step S12, that is, recognizing all the targets on the current image information, includes:

[0106] S121. Input the current image information into a preset contour recognition model for recognition to obtain the contour information of all candidate targets output by the contour recognition model;

[0107] S122. Input the current image information into a preset key feature recognition model for recognition to obtain the key feature information of all candidate targets output by the key feature recognition model;

[0108] S123. Recognize all the targets on the current image information according to the contour information and the key feature information.

[0109] It should be noted that when using recognition models to recognize human contours and human body parts in the prior art, it cannot be 100% accurate, and situations of mis-recognition or non-recognition often occur, which will have a great negative impact on subsequent human state perception.

[0110] Therefore, the embodiment of the present invention uses the combination of the recognition results of multiple recognition models to improve the accuracy of target recognition. The contour information of all candidate targets on the current image information is recognized according to a preset contour recognition model, and the key feature information of all candidate targets on the current image information is recognized according to a preset key feature recognition model, and the contour information and the key feature information are combined to recognize all the targets on the current image information.

[0111] Preferably, the contour information is the position information of the contour frame of the candidate target; the key feature information is the position information of the key feature points of the candidate target. Then, recognizing all the targets on the current image information according to the contour information and the key feature information includes:

[0112] According to the position information of the contour frame and the position information of the key feature points, when it is determined that any one of the key feature points is located within any one of the contour frames, the candidate target corresponding to the any one contour frame is taken as the target, so as to identify all the targets on the current image information.

[0113] As an example, if the target is a human body, then the key feature point is the human head.

[0114] Specifically, the embodiment of the present invention adopts a method of combining human body parts to determine the presence of a user. For example, the recognition accuracy of the human body contour and the head is relatively high. When the human body contour and the head exist simultaneously and the coordinate of the center point of the head is located within the human body contour, the two coordinates are combined and considered as one user. The recognition results of other parts within the user contour belong to the user, and finally all the recognition results located outside the user's human body contour are deleted. As Figure 9 shown, it is a schematic diagram of the principle of identifying targets in the embodiment of the present invention. For the two targets on the left, both the human body frame and the head can be recognized simultaneously and can be combined into two users. For the target on the right, only the head can be recognized, and the body contour is not recognized due to factors such as environmental airflow and distance, so this target is not regarded as a user.

[0115] It can be understood that in the actual application process, other parts of the human body can be determined as the key feature points. Similarly, for other types of targets, the key feature points can be determined according to the actual situation, which does not constitute a limitation to the present invention.

[0116] Preferably, the embodiment of the present invention is further implemented on the basis of the above embodiment. Step S12, that is, the step of identifying all the targets on the current image information, further includes:

[0117] According to the position information of all the targets, when it is determined that the distance between any two targets is less than or equal to a preset distance threshold, one of the any two targets is selected for retention, and the other target is deleted, so as to update all the targets identified on the current image information.

[0118] In the embodiment of the present invention, duplicate target processing is also performed. After all the targets are identified on the current image information, the position information of all the targets is extracted, and the recognition targets with high duplication are deleted by comparing the distances, and then the subsequent target tracking work is carried out to improve the accuracy of target recognition and target tracking.

[0119] See Figure 10 , which is a schematic flowchart of a target tracking method provided by the embodiment of the present invention. The embodiment of the present invention also provides a target tracking method, and the method includes steps S21 to S25:

[0120] S21. Obtain the current image information collected by a preset image acquisition module in real time;

[0121] S22. Identify all targets on the current image information;

[0122] S23. When a target is recognized for the first time, arrange and number the position information of the target on the current image information in a preset order to form a target sequence;

[0123] S24. When a target is recognized not for the first time, arrange the position information of the target on the current image information in a preset order;

[0124] S25. Match the arranged position information of the target on the current image information with the position information of the target sequence, and update the target sequence according to the matching result.

[0125] It should be noted that the target tracking method can be applied to an air conditioner, and the image acquisition module is arranged on the housing of the air conditioner.

[0126] By adopting the technical means of the embodiment of the present invention, the target is recognized and tracked by collecting the image information of the current environment in real time. After the target is recognized, the dynamic tracking of the target and the recording of the state change of the target are completed by matching with the historical position. The calculation is simple and the reliability is high. The calculation burden of the common tracking algorithm is significantly reduced, and the adverse effects caused by errors such as target misrecognition and under-recognition can be significantly reduced, so as to improve the accuracy of the operation control of the air conditioner and the user experience.

[0127] As a preferred embodiment, the preset order is the order of the abscissa of the position information of the target from small to large.

[0128] As a preferred embodiment, the matching the arranged position information of the target on the current image information with the position information of the target sequence and updating the target sequence according to the matching result includes:

[0129] Generate a first vector according to the arranged position information of the target on the current image information;

[0130] Generate a second vector according to the position information of the target in the current target sequence;

[0131] Construct an intersection over union matrix by calculating the intersection over union between the first vector and the second vector;

[0132] Update the number of the target corresponding to each position information in the first vector according to the intersection over union matrix to obtain an updated target sequence.

[0133] As a preferred embodiment, updating the numbers of the targets corresponding to each position information in the first vector according to the intersection-over-union ratio matrix to obtain an updated target sequence includes:

[0134] Determining the same target on the current image information as that in the target sequence according to the intersection-over-union ratio matrix;

[0135] Binding the numbers of the same target on the current image information to the same target in the target sequence;

[0136] Adding new numbers to the targets that are different from the target sequence on the current image information;

[0137] Until all the targets in the current image information are bound with numbers or new numbers are added, an updated target sequence is obtained.

[0138] As a preferred embodiment, updating the numbers of the targets corresponding to each position information in the first vector according to the intersection-over-union ratio matrix to obtain an updated target sequence includes:

[0139] Obtaining the target corresponding to the position information of the first vector corresponding to the maximum value in the intersection-over-union ratio matrix, denoted as the current target, and obtaining the target corresponding to the position information of the second vector corresponding to the maximum value in the intersection-over-union ratio matrix, denoted as the historical target;

[0140] Determining the current target and the historical target as the same target, and binding the number of the historical target to the current target as the number of the current target;

[0141] Setting the position information of the column vector and the row vector corresponding to the maximum value in the intersection-over-union ratio matrix to 0, and recalculating the intersection-over-union ratio matrix;

[0142] Until the numbers of all the targets corresponding to the position information in the first vector are bound to obtain an updated target sequence, or until there is no intersection-over-union ratio with a single maximum value or the maximum value is less than a preset intersection-over-union ratio threshold, adding new numbers to the targets corresponding to the remaining position information in the first vector to obtain an updated target sequence.

[0143] As a preferred embodiment, identifying all the targets on the current image information includes:

[0144] Inputting the current image information into a preset contour recognition model for recognition to obtain the contour information of all candidate targets output by the contour recognition model;

[0145] Input the current image information into a preset key feature recognition model for recognition to obtain the key feature information of all candidate targets output by the key feature recognition model;

[0146] Identify all targets on the current image information according to the contour information and the key feature information.

[0147] As a preferred embodiment, the contour information is the position information of the contour box of the candidate target; the key feature information is the position information of the key feature points of the candidate target;

[0148] Then, identifying all targets on the current image information according to the contour information and the key feature information includes:

[0149] According to the position information of the contour box and the position information of the key feature points, when it is determined that any one of the key feature points is located within any one of the contour boxes, the candidate target corresponding to the any one contour box is used as a target to identify all targets on the current image information.

[0150] As a preferred embodiment, identifying all targets on the current image information further includes:

[0151] According to the position information of all the targets, when it is determined that the distance between any two targets is less than or equal to a preset distance threshold, one of the any two targets is selected for retention and the other target is deleted to update all the targets identified on the current image information.

[0152] It should be noted that a target tracking method provided by an embodiment of the present invention has the same all process steps as those executed by a controller of an air conditioner in the above embodiment, and their working principles and beneficial effects correspond one by one, so they will not be elaborated herein.

[0153] See Figure 11 , which is a schematic structural diagram of a target tracking device provided by an embodiment of the present invention. An embodiment of the present invention also provides a target tracking device 30, including a processor 31, a memory 32, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the target tracking method as described in any one of the above embodiments

[0154] It should be noted that a target tracking device provided by an embodiment of the present invention is used to execute all process steps of a target tracking method in the above embodiment, and their working principles and beneficial effects correspond one by one, so they will not be elaborated herein.

[0155] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0156] The above is the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the technical field, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. An air conditioner, characterized in that, Comprising: A housing; An image acquisition module, disposed on the housing, for acquiring image information of the current indoor environment; A controller, configured to: Acquire in real time the current image information acquired by the image acquisition module; Identify all targets on the current image information; When a target is identified for the first time, arrange and number the position information of the target on the current image information in a preset order to form a target sequence; When a target is not identified for the first time, arrange the position information of the target on the current image information in a preset order; Match the position information of the target on the arranged current image information with the position information of the target sequence, and update the target sequence according to the matching result.

2. The air conditioner according to claim 1, wherein, The preset order is the order of the abscissas of the position information of the targets from small to large.

3. The air conditioner according to claim 1, characterized in that, The matching of the position information of the target on the arranged current image information with the position information of the target sequence and updating the target sequence according to the matching result includes: Generating a first vector according to the position information of the target on the arranged current image information; Generating a second vector according to the position information of the target in the current target sequence; Constructing an intersection over union matrix by calculating the intersection over union between the first vector and the second vector; Updating the number of the target corresponding to each position information in the first vector according to the intersection over union matrix to obtain an updated target sequence.

4. The air conditioner according to claim 3, characterized in that, The updating the number of the target corresponding to each position information in the first vector according to the intersection over union matrix to obtain an updated target sequence includes: Determining the same target on the current image information as the target in the target sequence according to the intersection over union matrix; Binding the number of the same target in the target sequence to the same target on the current image information; Adding a new number to the target on the current image information that is different from the target sequence; Until all targets in the current image information are bound with numbers or new numbers are added, an updated target sequence is obtained.

5. The air conditioner according to claim 3, characterized in that, The updating the number of the target corresponding to each position information in the first vector according to the intersection over union matrix to obtain an updated target sequence includes: Obtaining the target corresponding to the position information of the first vector corresponding to the maximum value in the intersection over union matrix, denoted as the current target, and obtaining the target corresponding to the position information of the second vector corresponding to the maximum value in the intersection over union matrix, denoted as the historical target; Determining that the current target and the historical target are the same target, and binding the number of the historical target to the current target as the number of the current target; Setting the position information of the column vector and the row vector corresponding to the maximum value in the intersection over union matrix to 0, and recalculating the intersection over union matrix; Until all targets corresponding to the position information in the first vector are bound with numbers, an updated target sequence is obtained, or until there is no single maximum intersection over union or the maximum value is less than a preset intersection over union threshold, new numbers are added to the targets corresponding to the remaining position information in the first vector to obtain an updated target sequence.

6. The air conditioner according to claim 1, characterized in that, Identifying all targets on the current image information includes: Inputting the current image information into a preset contour recognition model for recognition to obtain the contour information of all candidate targets output by the contour recognition model; Inputting the current image information into a preset key feature recognition model for recognition to obtain the key feature information of all candidate targets output by the key feature recognition model; Identifying all targets on the current image information according to the contour information and the key feature information.

7. The air conditioner according to claim 6, characterized in that, The contour information is the position information of the contour frame of the candidate target; the key feature information is the position information of the key feature points of the candidate target; Then, identifying all targets on the current image information according to the contour information and the key feature information includes: According to the position information of the contour frame and the position information of the key feature points, when it is determined that any one of the key feature points is located within any one of the contour frames, the candidate target corresponding to the any one contour frame is used as a target to identify all targets on the current image information.

8. The air conditioner according to claim 6 or 7, characterized in that, Identifying all targets on the current image information further includes: According to the position information of all the targets, when it is determined that the distance between any two targets is less than or equal to a preset distance threshold, one target is selected and retained from the any two targets, and the other target is deleted to update all the targets identified on the current image information.

9. A target tracking method, characterized in that, Including: Real-time acquiring the current image information acquired by a preset image acquisition module; Identifying all targets on the current image information; When targets are recognized for the first time, arranging and numbering the position information of the targets on the current image information in a preset order to form a target sequence; When targets are not recognized for the first time, arranging the position information of the targets on the current image information in a preset order; Matching the arranged position information of the targets on the current image information with the position information of the target sequence, and updating the target sequence according to the matching result.

10. A target tracking device, characterized in that, Including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the target tracking method according to any one of claims 9 is implemented.

Images