A heat control method and system for a heating ultrasonic probe

Abstract

The present invention discloses a method and system for controlling the heat of a heating ultrasonic probe. Through a first temperature-time convolutional network, using a temperature-delay image representing the temperatures corresponding to the start power-on time point and all the stop power-on time points, the relationship between the temperatures at different stop power-on time points and the delay of the stop power-on time points is found as a whole. Then, a time-temperature curve is constructed and a three-dimensional time-temperature image is obtained. By convolving the three-dimensional time-temperature image through a second temperature-time convolutional network, the change in temperature at time points with the same time length from the stop power-on time point is found. And by using the differences in the coefficients of the time-temperature curve, the change in temperature at multiple stop surrounding time points between two stop power-on time points is found.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular, to a heat control method and system for a heating ultrasonic probe. Background Art

[0002] Currently, since the handheld ultrasonic probe used generates heat during use, and excessive heat can damage the tissues and bacteria in the surgical area, a thermistor needs to be installed in the handheld ultrasonic probe to monitor the heat during the operation of the ultrasonic probe.

[0003] However, there is a delay between the heating time length and the heat reached, and the current heat at which heating stops is not the highest temperature of the handheld ultrasonic probe, but the highest temperature of the handheld ultrasonic probe appears some time after heating stops.

[0004] Therefore, it is necessary to find the correlation between the two. Summary of the Invention

[0005] The purpose of the present invention is to provide a heat control method and system for a heating ultrasonic probe to solve the above problems existing in the prior art.

[0006] In a first aspect, an embodiment of the present invention provides a heat control method for a heating ultrasonic probe, including:

[0007] Obtain the start power-on time point and multiple stop power-on time points; the start power-on time point represents the time point when power-on starts; the stop power-on time point represents the time point when power-on stops;

[0008] Based on multiple stop power-on time points, obtain corresponding multiple probe time-temperature images; the probe time-temperature image includes the temperatures at multiple time points around the stop; the time points around the stop include n time points earlier than the stop power-on time point and n time points later than the stop power-on time point;

[0009] Based on multiple probe time-temperature images, detect the highest temperature at the stop power-on time point and the corresponding time points around the stop to obtain a temperature delay image;

[0010] Through a first temperature-time convolutional network, based on the temperature delay image, obtain a first temperature delay feature; the first temperature delay feature is used to detect the delay degree of the temperatures at multiple stop power-on time points;

[0011] Based on multiple probe time-temperature images, judge the temperature changes corresponding to different stop power-on time points to obtain a second temperature delay feature;

[0012] Input the first temperature delay feature and the second temperature delay feature into a prediction network to obtain a predicted power-off time point; the predicted power-off time point represents the earliest predicted power-off time point that can reach a preset temperature.

[0013] Send a signal to stop power supply at the predicted power-off time point.

[0014] Optionally, obtaining the first temperature delay feature through the first temperature-time convolutional network based on the temperature delay image includes:

[0015] The first temperature-time convolutional network includes a two-dimensional convolutional kernel of p*((2n + 1)*2);

[0016] p represents the width of the temperature delay image, (2n + 1) represents the time length between the power-off time point and the corresponding multiple surrounding power-off time points; ((2n + 1)*2) represents using the two-dimensional convolutional kernel to find the delay change between the temperatures of two power-off time points and the corresponding highest temperature;

[0017] With a stride of ((2n + 1)*2), perform convolution on the temperature delay image with the two-dimensional convolutional kernel included in the first temperature-time convolutional network to obtain a first detection feature vector;

[0018] Input the first detection feature vector into a first neural network to detect the feature of the temperature change between two power-off time points, and obtain the first temperature delay feature.

[0019] Optionally, obtaining the corresponding multiple probe time-temperature images based on multiple power-off time points includes:

[0020] Obtain multiple surrounding power-off time points; the surrounding power-off time points include n time points earlier than the power-off time point and n time points later than the power-off time point; the time length between any two adjacent surrounding power-off time points is the same;

[0021] Obtain the temperatures corresponding to the multiple surrounding power-off time points as probe temperatures;

[0022] Construct a time-temperature coordinate axis with the time point as the abscissa and the probe temperature as the ordinate; the abscissa of the origin of the time-temperature coordinate axis represents the power-on start time point; the ordinate of the origin of the time-temperature coordinate axis represents a temperature of 0;

[0023] With the abscissa of the time-temperature coordinate axis as the length and the ordinate as the width, plot the surrounding power-off time points and the corresponding probe temperatures as a probe time-temperature image; the length of the probe time-temperature image represents the time length between 3 adjacent power-off time points; the width of the probe time-temperature image represents 0 to 100 degrees Celsius.

[0024] The probe time-temperature image is a dot plot.

[0025] Optionally, judging the temperature changes corresponding to different power-off time points based on multiple probe time-temperature images to obtain a second temperature delay feature, including:

[0026] Construct a curve with the power-off time points and multiple surrounding time points in the probe time-temperature image and the corresponding temperatures to obtain a time-temperature curve;

[0027] Multiple probe time-temperature images correspond to obtaining multiple time-temperature curves;

[0028] Stack multiple probe time-temperature images in order from the earliest to the latest power-off time point to obtain a three-dimensional time-temperature image;

[0029] Based on the three-dimensional time-temperature image, obtain a first change feature through a second temperature-time convolutional network;

[0030] Detect coefficient changes based on the multiple time-temperature curves to obtain a second change feature;

[0031] Obtain a second temperature delay feature according to the first change feature and the second change feature.

[0032] Optionally, the obtaining of the first change feature through a second temperature-time convolutional network based on the three-dimensional time-temperature image includes:

[0033] The second temperature-time convolutional network includes a three-dimensional temperature-time convolutional network and a two-dimensional temperature-time convolutional network;

[0034] The three-dimensional temperature-time convolutional network includes a three-dimensional convolutional kernel of p*2*2; p represents the height of the three-dimensional time-temperature image;

[0035] In the width direction of the three-dimensional time-temperature image, with a step size of 1, convolve the three-dimensional convolutional kernel of the three-dimensional time-temperature image on the three-dimensional time-temperature image to detect the temperature changes of different starting temperatures at the same power-on time, and obtain a first feature image; the width of the first feature image corresponds to the width of the three-dimensional time-temperature image; the length of the first feature image corresponds to the length of the three-dimensional time-temperature image;

[0036] Based on the first feature image, obtain a first change feature through a two-dimensional temperature-time convolutional network; the first change feature represents the feature change of adjacent power-off time points.

[0037] Optionally, the obtaining of the first change feature through a two-dimensional temperature-time convolutional network based on the first feature image includes:

[0038] The two-dimensional temperature-time convolutional network includes a two-dimensional convolutional kernel of m*2; m represents the width of the first feature image;

[0039] In the length direction of the first feature image, with a stride of 1, the two-dimensional convolutional kernel of the two-dimensional temperature-time convolutional network is convolved on the first feature image to obtain a first change feature; the first change feature is a one-dimensional vector.

[0040] Optionally, the detecting coefficient change based on the multiple time-temperature curves to obtain a second change feature includes:

[0041] Input the coefficients of the time-temperature curves into the time convolutional network in sequence according to the power-off time points from early to late to obtain a second change feature.

[0042] Optionally, the detecting the highest temperature at the power-off time point and the corresponding surrounding time points based on multiple probe time-temperature images to obtain a temperature delay image includes:

[0043] In the probe time-temperature image, detect the highest temperature value as the highest temperature; multiple probe time-temperature images correspond to multiple highest temperatures;

[0044] In the probe time-temperature image, take the time point corresponding to the highest temperature as the highest temperature time point;

[0045] Take the temperature corresponding to the power-off time point of the probe time-temperature image as the power-off temperature;

[0046] Take the power-off time point, power-off temperature, highest temperature time point, and highest temperature corresponding to the probe time-temperature image as a time-temperature set;

[0047] Multiple probe time-temperature images correspond to obtain multiple time-temperature sets;

[0048] Based on multiple time-temperature sets, obtain a temperature delay image.

[0049] Optionally, the obtaining a temperature delay image based on multiple time-temperature sets includes:

[0050] Obtain the latest power-off time point; the latest power-off time point represents the power-off time point that is later than other power-off time points among multiple power-off time points;

[0051] Obtain a complete time image; the length of the complete time image represents multiple power-off time points from the power-on time point to the latest power-off time point;

[0052] Mark the points formed by the power-off time point and the power-off temperature in the time-temperature set on the complete time image to obtain the stop position;

[0053] Mark the points formed by the highest power-on time point and the highest power-on temperature in the time-temperature set on the complete time image to obtain the highest position; Multiple time-temperature sets correspond to obtaining multiple stop positions and multiple highest positions;

[0054] Connect the stop positions and the corresponding highest positions in the complete time image with straight lines to obtain the temperature delay image.

[0055] In a second aspect, an embodiment of the present invention provides a heat control system for a heating ultrasonic probe, including:

[0056] An acquisition module for acquiring the start power-on time point and multiple stop power-on time points; The start power-on time point represents the time point when power-on starts; The stop power-on time point represents the time point when power-on stops; Based on multiple stop power-on time points, obtain corresponding multiple probe time-temperature images; The probe time-temperature image includes the temperatures of multiple time points around the stop; The time points around the stop include n time points earlier than the stop power-on time point and n time points later than the stop power-on time point;

[0057] A temperature delay image acquisition module for detecting the highest temperature of the stop power-on time point and the corresponding time points around the stop based on multiple probe time-temperature images to obtain a temperature delay image;

[0058] A first temperature delay detection module for obtaining a first temperature delay feature based on the temperature delay image through a first temperature-time convolutional network; The first temperature delay feature is used to detect the delay degree of the temperatures at multiple stop power-on time points;

[0059] A second temperature delay detection module for judging the temperature changes corresponding to different stop power-on time points based on multiple probe time-temperature images to obtain a second temperature delay feature;

[0060] A prediction module for inputting the first temperature delay feature and the second temperature delay feature into a prediction network to obtain a predicted stop power-on time point; The predicted stop power-on time point represents the earliest stop power-on time point predicted to reach a preset temperature;

[0061] A reminder module for sending a power-off signal at the predicted stop power-on time point.

[0062] Compared with the prior art, the embodiments of the present invention achieve the following beneficial effects:

[0063] An embodiment of the present invention also provides a heat control method and system for a heating ultrasonic probe.

[0064] In the present invention, there is a delay between the time when heating stops and the highest temperature reached by the handheld ultrasonic probe.

[0065] Through the first temperature-time convolutional network, using a temperature-delay image representing the temperatures corresponding to the start power-on time point and all stop power-on time points, the relationship between the temperatures at different stop power-on time points and the delay of the stop power-on time points is found as a whole.

[0066] Then, a time-temperature curve is constructed and a three-dimensional time-temperature image is obtained.

[0067] By convolving the three-dimensional time-temperature image through the second temperature-time convolutional network, the temperature changes at time points with the same time length from the stop power-on time point are found.

[0068] And by using the differences in the coefficients of the time-temperature curve, the temperature changes at multiple stop surrounding time points between two stop power-on time points are found.

[0069] The technical effect of being able to find the correlation relationship between the stop power-on time point and the temperature, so as to accurately predict the time point when the preset temperature is reached is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 is a flowchart of a heat control method for a heating ultrasonic probe provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0071] The present invention will be described in detail below with reference to the accompanying drawings.

[0072] Embodiment 1

[0073] As Figure 1 shown, an embodiment of the present invention provides a heat control method for a heating ultrasonic probe, and the method includes:

[0074] S101: Obtain the start power-on time point and multiple stop power-on time points; the start power-on time point represents the time point when power-on starts; the stop power-on time point represents the time point when power-on stops.

[0075] Among them, the time period length between any two adjacent stop power-on time points is the same.

[0076] Among them, the power consumption of multiple power-on time points is the same.

[0077] S102: Obtain a plurality of probe time-temperature images corresponding to a plurality of power-off time points; the probe time-temperature images include temperatures at a plurality of time points around the power-off; the time points around the power-off include n time points earlier than the power-off time point and n time points later than the power-off time point.

[0078] S103: Based on the plurality of probe time-temperature images, detect the highest temperatures at the power-off time point and the corresponding time points around the power-off to obtain a temperature delay image.

[0079] S104: Through a first temperature-time convolutional network, based on the temperature delay image, obtain a first temperature delay feature; the first temperature delay feature is used to detect the delay degree of the temperatures at a plurality of power-off time points.

[0080] S105: Based on the plurality of probe time-temperature images, judge the temperature changes corresponding to different power-off time points to obtain a second temperature delay feature.

[0081] S106: Input the first temperature delay feature and the second temperature delay feature into a prediction network to obtain a predicted power-off time point; the predicted power-off time point represents the earliest predicted power-off time point that can reach a preset temperature.

[0082] Wherein, in this embodiment, the prediction network is a fully connected neural network (FCNN).

[0083] Wherein, the prediction network is a network trained with historical data.

[0084] S107: Send a power-off signal at the predicted power-off time point.

[0085] Establish a relationship between the power consumption at the plurality of times and the probe temperature, and stop at a certain temperature; (the temperature will increase and decrease after a delay within a certain range)

[0086] Optionally, the step of obtaining a first temperature delay feature through a first temperature-time convolutional network based on the temperature delay image includes:

[0087] The first temperature-time convolutional network includes a two-dimensional convolutional kernel of p*((2n + 1)*2).

[0088] Wherein, p and n are positive integers.

[0089] Let \(p\) represent the width of the temperature delay image, and \((2n + 1)\) represent the time length between the power-off time point and the corresponding multiple surrounding time points; \(((2n + 1)\times2)\) represents the change in delay between the temperatures at two power-off time points and the corresponding highest temperature found using a two-dimensional convolution kernel.

[0090] With a step size of \(((2n + 1)\times2)\), convolve the two-dimensional convolution kernel included in the first temperature-time convolution network on the temperature delay image to obtain a first detection feature vector.

[0091] Through the above method, the special two-dimensional convolution kernel included in the first temperature-time convolution network can convert the image into a one-dimensional vector.

[0092] Input the first detection feature vector into the first neural network to detect the features of the temperature change at two power-off time points, and obtain a first temperature delay feature.

[0093] Among them, the first neural network is a fully connected neural network (FCNN).

[0094] Optionally, the obtaining of the corresponding multiple probe time-temperature images based on multiple power-off time points includes:

[0095] Obtain multiple surrounding time points; the surrounding time points include \(n\) time points earlier than the power-off time point and \(n\) time points later than the power-off time point; the time length between any two adjacent surrounding time points is the same;

[0096] Obtain the temperatures corresponding to the multiple surrounding time points as probe temperatures;

[0097] Construct a time-temperature coordinate axis with the time point as the abscissa and the probe temperature as the ordinate; the abscissa of the origin of the time-temperature coordinate axis represents the power-on start time point; the ordinate of the origin of the time-temperature coordinate axis represents a temperature of 0;

[0098] With the abscissa of the time-temperature coordinate axis as the length and the ordinate as the width, plot the surrounding time points and the corresponding probe temperatures to form a probe time-temperature image; the length of the probe time-temperature image represents the time length between three adjacent power-off time points; the width of the probe time-temperature image represents 0 to 100 degrees Celsius.

[0099] The probe time-temperature image is a dot plot.

[0100] Among them, the lower left corner of the probe time-temperature image is the origin.

[0101] Among them, in this embodiment, the probe time-temperature image has a size of 512*512. The coordinates 0, 1, …, 511, 512 of the coordinate axis respectively correspond to each pixel point of the probe time-temperature image. The length of the probe time-temperature image is evenly divided into 2n + 1 points. For example, when n = 8, taking the abscissa 0 as the starting point, and rounding down (512 / 17) to 30, the points on the coordinate axis are respectively: (0,0), (30,y1), (60,y2), (90,y3), (120,y4), (150,y5), (180,y6), (210,y7), (240,y8), (270,y9), (300,y10), (330,y11), (360,y12), (390,y13), (420,y14), (450y15), (480,y16). 0, y1, y2, y3, y4, y5, y6, y7, y8, y9, y10, y11, y12, y13, y14, y15, y16 represent the temperatures corresponding to multiple stop surrounding time points and stop power-on time points, and y8 is the temperature corresponding to the stop power-on time point.

[0102] Among them, the probe time-temperature image is a binary image.

[0103] Optionally, the judging the change of the temperature corresponding to different stop power-on time points based on multiple probe time-temperature images to obtain the second temperature delay feature includes:

[0104] Construct a curve with the stop power-on time point and multiple stop surrounding time points in the probe time-temperature image and the corresponding temperatures to obtain a time-temperature curve.

[0105] Among them, in this embodiment, the curve image is constructed by using polynomial interpolation with multiple points represented by the stop power-on time point and multiple stop surrounding time points in the probe time-temperature image and the corresponding temperatures.

[0106] Multiple probe time-temperature images correspond to obtain multiple time-temperature curves;

[0107] Stack multiple probe time-temperature images in ascending order of the stop power-on time point to obtain a three-dimensional time-temperature image.

[0108] Among them, the length of the three-dimensional time-temperature image is equal to the length of the curve image, the height of the three-dimensional time-temperature image is equal to the height of the curve image, and the width of the three-dimensional time-temperature image is equal to the sum of the widths of multiple curve images.

[0109] Based on the three-dimensional time-temperature image, obtain the first change feature through the second temperature-time convolutional network.

[0110] Based on the multiple time-temperature curves, detect the coefficient change to obtain a second change feature.

[0111] Obtain a second temperature delay feature according to the first change feature and the second change feature.

[0112] Optionally, the obtaining of the first change feature based on the three-dimensional time-temperature image through the second temperature-time convolutional network includes:

[0113] The second temperature-time convolutional network includes a three-dimensional temperature-time convolutional network and a two-dimensional temperature-time convolutional network.

[0114] Wherein, in this embodiment, the three-dimensional temperature-time convolutional network is a three-dimensional convolutional neural network (3D Convolutional Neural Networks, CNN). The two-dimensional temperature-time convolutional network is a convolutional neural network (Convolutional Neural Networks, CNN).

[0115] The three-dimensional temperature-time convolutional network includes a three-dimensional convolutional kernel of p*2*2; p represents the height of the three-dimensional time-temperature image;

[0116] In the width direction of the three-dimensional time-temperature image, with a stride of 1, convolve the three-dimensional convolutional kernel of the three-dimensional time-temperature image on the three-dimensional time-temperature image to detect the temperature change at the same energization time for different starting temperatures, and obtain a first feature image; the width of the first feature image corresponds to the width of the three-dimensional time-temperature image; the length of the first feature image corresponds to the length of the three-dimensional time-temperature image.

[0117] Based on the first feature image, obtain a first change feature through the two-dimensional temperature-time convolutional network; the first change feature represents the feature change of adjacent power-off time points.

[0118] Optionally, the obtaining of the first change feature based on the first feature image through the two-dimensional temperature-time convolutional network includes:

[0119] The two-dimensional temperature-time convolutional network includes a two-dimensional convolutional kernel of m*2; m represents the width of the first feature image.

[0120] Wherein, m is a positive integer.

[0121] In the length direction of the first feature image, with a stride of 1, convolve the two-dimensional convolutional kernel of the two-dimensional temperature-time convolutional network on the first feature image to obtain a first change feature; the first change feature is a one-dimensional vector.

[0122] Optionally, the detecting coefficient change based on the multiple time-temperature curves to obtain a second change feature includes:

[0123] Input the coefficients of the time-temperature curves into a temporal convolutional network in order from the earliest power-off time point to the latest, to obtain a second change feature.

[0124] Optionally, the detecting the highest temperature at the power-off time point and the corresponding surrounding time points based on multiple probe time-temperature images to obtain a temperature delay image includes:

[0125] In the probe time-temperature image, detect the highest temperature value as the highest temperature; there are multiple highest temperatures corresponding to multiple probe time-temperature images;

[0126] In the probe time-temperature image, take the time point corresponding to the highest temperature as the highest temperature time point;

[0127] Take the temperature corresponding to the power-off time point of the probe time-temperature image as the power-off temperature;

[0128] Take the power-off time point, power-off temperature, highest temperature time point, and highest temperature corresponding to the probe time-temperature image as a time-temperature set;

[0129] Multiple probe time-temperature images correspond to obtain multiple time-temperature sets;

[0130] Based on multiple time-temperature sets, obtain a temperature delay image.

[0131] Optionally, the obtaining a temperature delay image based on multiple time-temperature sets includes:

[0132] Obtain the latest power-off time point; the latest power-off time point represents the power-off time point that is later than other power-off time points among multiple power-off time points;

[0133] Obtain a complete time image; the length of the complete time image represents multiple power-off time points from the start power-on time point to the latest power-off time point.

[0134] Wherein, in this embodiment, if it is divided into 2n time points between two adjacent power-off time points and the number of power-off time points is m, then the complete time image is divided into 2n*m points to represent the power-off time points and the surrounding time points.

[0135] Mark the points composed of the power-off time point and power-off temperature in the time-temperature set in the complete time image to obtain a stop position;

[0136] Mark the point formed by the highest power-on time point and the highest power-on temperature in the time-temperature set in the complete time image to obtain the highest position; multiple time-temperature sets correspond to obtain multiple stop positions and multiple highest positions.

[0137] Among them, the highest power-on time point is one of the stop surrounding time points among multiple stop surrounding time points.

[0138] Connect the stop positions and the corresponding highest positions in the complete time image with straight lines to obtain the temperature delay image.

[0139] Embodiment 2

[0140] Based on the above heat control method for a heating ultrasonic probe, an embodiment of the present invention further provides a heat control system for a heating ultrasonic probe. The system includes an acquisition module, a temperature delay image acquisition module, a first temperature delay detection module, a second temperature delay detection module, a prediction module, and a reminder module.

[0141] The acquisition module is used to acquire the start power-on time point and multiple stop power-on time points; the start power-on time point represents the time point when power-on starts; the stop power-on time point represents the time point when power-on stops; based on multiple stop power-on time points, obtain the corresponding multiple probe time-temperature images; the probe time-temperature image includes the temperatures of multiple stop surrounding time points; the stop surrounding time points include n time points earlier than the stop power-on time point and n time points later than the stop power-on time point;

[0142] The temperature delay image acquisition module is used to detect the highest temperature of the stop power-on time point and the corresponding stop surrounding time points based on multiple probe time-temperature images to obtain the temperature delay image;

[0143] The first temperature delay detection module is used to obtain the first temperature delay feature based on the temperature delay image through the first temperature-time convolutional network; the first temperature delay feature is used to detect the delay degree of the temperatures of multiple stop power-on time points;

[0144] The second temperature delay detection module is used to judge the temperature changes corresponding to different stop power-on time points based on multiple probe time-temperature images to obtain the second temperature delay feature;

[0145] The prediction module is used to input the first temperature delay feature and the second temperature delay feature into the prediction network to obtain the predicted stop power-on time point; the predicted stop power-on time point represents the earliest stop power-on time point predicted to reach the preset temperature;

[0146] The reminder module is used to send a stop power-on signal at the predicted stop power-on time point.

[0147] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system, or other device.

[0148] Various general-purpose systems may also be used in conjunction with the teachings based hereon. The structure required to construct such systems will be apparent from the above description.

[0149] Furthermore, the present invention is not directed to any particular programming language.

[0150] It should be understood that the teachings of the present invention described herein can be implemented in various programming languages, and the description of a particular language above is for the purpose of disclosing the best mode of the present invention.

Claims

1. A method for controlling the heat of a heated ultrasonic probe, characterized in that: include: Acquire a power-on start time point and a plurality of power-off stop time points; the power-on start time point indicates a time point at which power is started; the power-off stop time point indicates a time point at which power is stopped; Based on multiple power-off time points, a corresponding multiple probe time-temperature images are acquired; the probe time-temperature images include temperatures of multiple stop-around time points; the stop-around time points include n time points earlier than the power-off time point and n time points later than the power-off time point; Based on multiple probe time-temperature images, the highest temperature at the time point of power-off and the corresponding time points around the power-off is detected to obtain a temperature time-lapse image; A first temperature delay feature is obtained based on the temperature delay image through a first temperature-time convolution network; the first temperature delay feature is used to detect the degree of delay of the temperature at multiple power-off time points; Based on multiple probe time-temperature images, determine the temperature changes corresponding to different power-off time points to obtain a second temperature delay feature; The first temperature delay feature and the second temperature delay feature are input into a prediction network to obtain a predicted power-off time point; the predicted power-off time point represents the earliest predicted power-off time point that can reach a preset temperature; A signal to stop energizing is sent at the predicted energizing stop time point.

2. A method for controlling the heat of a heated ultrasonic probe according to claim 1, characterized in that: The first temperature time-delay feature is obtained based on the temperature time-delay image through the first temperature time convolution network, including: The first temperature-time convolution network includes a two-dimensional convolution kernel of p*((2n+1)*2); p represents the width of the temperature delay image, (2n+1) represents the time length between the power-off time point and the corresponding multiple stop surrounding time points; ((2n+1)*2) represents the change in the delay between the temperature at two power-off time points and the corresponding highest temperature found by using a two-dimensional convolution kernel; With a step size of ((2n+1)*2), convolving the two-dimensional convolution kernel included in the first temperature-time convolution network on the temperature delay image to obtain a first detection feature vector; The first detection feature vector is input into a first neural network to detect the characteristics of the temperature change at two power-off time points to obtain a first temperature delay feature.

3. A method for controlling the heat of a heated ultrasonic probe according to claim 1, characterized in that: The step of acquiring corresponding multiple probe time-temperature images based on multiple power-off time points includes: Acquire multiple time points around the stop; the time points around the stop include n time points earlier than the power-off time point and n time points later than the power-off time point; the time lengths of any two adjacent time points around the stop are the same; Obtain the temperatures corresponding to multiple stop-around time points as probe temperatures; A time-temperature coordinate axis is constructed with the time point as the abscissa and the probe temperature as the ordinate; the abscissa of the origin of the time-temperature coordinate axis represents the time point when power is turned on; the ordinate of the origin of the time-temperature coordinate axis represents the temperature as 0; With the horizontal axis of the time-temperature coordinate axis as the length and the vertical axis as the width, the probe time-temperature image is drawn by taking the stop surrounding time points and the corresponding probe temperatures; the length of the probe time-temperature image represents the time length between three adjacent stop power-on time points; the width of the probe time-temperature image represents 0 to 100 degrees Celsius; The probe time-temperature image is a point image.

4. A method for controlling the heat of a heated ultrasonic probe according to claim 1, characterized in that: The method of determining the temperature change corresponding to different power-off time points based on the multiple probe time-temperature images to obtain the second temperature delay feature includes: The time point of power-off and multiple time points around power-off in the probe time-temperature image are used to construct a curve with the corresponding temperature to obtain a time-temperature curve; Multiple probe time-temperature images correspond to multiple time-temperature curves; According to the time point of power off from early to late, multiple probe time-temperature images are superimposed to obtain a three-dimensional time-temperature image; Obtaining a first change feature based on the three-dimensional time-temperature image through a second temperature-time convolutional network; Based on the multiple time-temperature curves, detect the coefficient change to obtain a second change feature; A second temperature delay characteristic is obtained according to the first change characteristic and the second change characteristic.

5. A method for controlling the heat of a heated ultrasonic probe according to claim 4, characterized in that: The first change feature is obtained based on the three-dimensional time-temperature image through the second temperature-time convolution network, including: The second temperature-time convolution network includes a three-dimensional temperature-time convolution network and a two-dimensional temperature-time convolution network; The three-dimensional temperature-time convolution network includes a three-dimensional convolution kernel of p*2*2; p represents the height of the three-dimensional time-temperature image; In the width direction of the three-dimensional time-temperature image, the three-dimensional convolution kernel of the three-dimensional time-temperature image is convolved on the three-dimensional time-temperature image with a step size of 1, and the temperature change of different starting temperatures when the power is turned on for the same time is detected to obtain a first characteristic image; the width of the first characteristic image corresponds to the width of the three-dimensional time-temperature image; the length of the first characteristic image corresponds to the length of the three-dimensional time-temperature image; Based on the first feature image, a first change feature is obtained through a two-dimensional temperature-time convolution network; the first change feature represents the feature of the feature change of adjacent power-off time points.

6. A method for controlling the heat of a heated ultrasonic probe according to claim 5, characterized in that: The obtaining of a first change feature based on the first feature image through a two-dimensional temperature-time convolution network includes: The two-dimensional temperature-time convolution network includes a two-dimensional convolution kernel of m*2; m represents the width of the first feature image; In the long direction of the first feature image, with a step size of 1, the two-dimensional convolution kernel of the two-dimensional temperature-time convolution network is convolved on the first feature image to obtain a first change feature; the first change feature is a one-dimensional vector.

7. A method for controlling the heat of a heated ultrasonic probe according to claim 4, characterized in that: The detecting coefficient change based on the multiple time-temperature curves to obtain a second change feature includes: According to the power-off time points from early to late, the coefficients of the time-temperature curve are sequentially input into the time convolution network to obtain the second change feature.

8. The method for controlling the heat of a heated ultrasonic probe according to claim 1, characterized in that: The method of detecting the highest temperature at the power-off time point and the corresponding surrounding time points based on multiple probe time-temperature images to obtain a temperature delay image includes: In the probe time-temperature image, the highest detected temperature value is taken as the highest temperature; multiple probe time-temperature images correspond to multiple highest temperatures; In the probe time-temperature image, the time point corresponding to the highest temperature is taken as the highest temperature time point; The temperature corresponding to the power-off time point corresponding to the probe time-temperature image is used as the power-off temperature; The power-off time point, power-off temperature, maximum temperature time point and maximum temperature corresponding to the probe time-temperature image are taken as a time-temperature set; Multiple probe time-temperature images correspond to obtaining multiple time-temperature sets; Based on multiple time-temperature sets, a temperature time-lapse image is obtained.

9. A method for controlling the heat of a heated ultrasonic probe according to claim 8, characterized in that: The step of obtaining a temperature time-lapse image based on multiple time-temperature sets includes: Obtaining the latest power-off time point; the latest power-off time point represents a power-off time point that is later than other power-off time points among the multiple power-off time points; Acquire a complete time image; the length of the complete time image represents multiple de-energization time points from the start time point of energization to the latest de-energization time point; Mark the point consisting of the power-off time point and the power-off temperature in the time-temperature set in the complete time image to obtain the stop position; The point consisting of the highest power-on time point and the highest power-on temperature in the time-temperature set is marked in the complete time image to obtain the highest position; multiple stop positions and multiple highest positions are obtained corresponding to multiple time-temperature sets; The stop position and the corresponding highest position in the complete time image are connected with a straight line to obtain a temperature time-lapse image.

10. A heat control system for a heated ultrasonic probe, characterized in that: include: An acquisition module is used to acquire a power-on start time point and a plurality of power-off stop time points; the power-on start time point indicates a power-on start time point; the power-off stop time point indicates a power-off stop time point; based on the plurality of power-off stop time points, a plurality of corresponding probe time-temperature images are acquired; the probe time-temperature images include temperatures at a plurality of stop surrounding time points; the stop surrounding time points include n time points earlier than the power-off stop time point and n time points later than the power-off stop time point; A temperature time-lapse image acquisition module is used to detect the highest temperature at the power-off time point and the corresponding surrounding time points based on multiple probe time temperature images to obtain a temperature time-lapse image; A first temperature delay detection module, configured to obtain a first temperature delay feature based on the temperature delay image through a first temperature time convolution network; the first temperature delay feature is used to detect the degree of temperature delay at multiple power-off time points; A second temperature delay detection module is used to determine the temperature change corresponding to different power-off time points based on multiple probe time-temperature images to obtain a second temperature delay feature; A prediction module, configured to input the first temperature delay feature and the second temperature delay feature into a prediction network to obtain a predicted power-off time point; the predicted power-off time point represents the earliest predicted power-off time point that can reach a preset temperature; The reminder module is used to send a power-off signal at the predicted power-off time point.

Images