State transition based thermal imaging anti-burnout apparatus and method
By using a state-transfer-based thermal imaging method, combined with gradient detection and watchdog detection, and utilizing a miniature thermal imaging camera and thermocouple valve, the difficulties of existing anti-dry-burning stoves in terms of threshold setting and pot material differences have been solved, realizing low-cost and convenient dry-burning warning and automatic valve shut-off functions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHENGDU QIANJIA TECH CO LTD
- Filing Date
- 2023-12-19
- Publication Date
- 2026-06-12
AI Technical Summary
Existing anti-dry-burning stoves face difficulties in setting thresholds and considering differences in pot materials, resulting in an inability to accurately detect dry-burning behavior. They are also costly and complex to install.
A state-transition-based thermal imaging method is adopted. By combining low-frequency, high-frequency and gradient detection modes with watchdog detection, a miniature thermal imaging camera is used to collect thermal images of the stove to accurately perceive dry burning behavior. Automatic valve closure is achieved through a control module and thermocouple valve.
It achieves accurate detection of dry burning behavior, is low in cost and easy to install, and avoids the problems of frequent false alarms and failure to provide timely warnings.
Smart Images
Figure CN117739371B_ABST
Abstract
Claims
1. A thermal imaging-based method for preventing dry burning based on state transition, characterized in that: Includes the following steps: Step 1: Under low-frequency sampling conditions, thermal images are acquired using thermal imaging in temperature tracking mode. The highest temperature value of the thermal images is counted. If the highest temperature value is within the first temperature threshold range, thermal images are acquired again in temperature tracking mode; otherwise, proceed to step 2. Step 2: If the highest temperature value of the thermal image is within the second temperature threshold range, switch to high-frequency sampling mode; in high-frequency sampling mode, when the highest temperature value of the acquired thermal image is within the first threshold range, return to the temperature tracking mode of Step 1; when the highest temperature value is within the second temperature threshold range, continue with the temperature tracking mode of Step 2; when the highest temperature value is within the third temperature threshold range, enter the gradient detection mode of Step 3; when the highest temperature value is within the fourth temperature threshold range, enter the watchdog detection mode of Step 4. Step 3: If the highest temperature value of the thermal image is within the range of the third temperature threshold, switch to gradient detection mode to determine whether the highest temperature value of a set of thermal images is monotonically increasing. If it is not monotonically increasing, perform order-preserving transformation to make it monotonically increasing. Then compare the temperature increase of this group with the set threshold. If the temperature increase is greater than the set threshold, it is judged that the dry burning state is approaching and a warning is issued. Otherwise, return to step 2. Step 4: If the highest temperature value of the thermal image is within the range of the fourth temperature threshold, switch to watchdog detection mode. After multiple samplings, determine whether it is in a dry-burning state. If it is in a dry-burning state, issue an alert and close the valve.
2. The thermal imaging anti-dry-burning method based on state transition according to claim 1, characterized in that: Step 1 specifically includes the following steps: Set a first temperature threshold t1, a second temperature threshold t2, a third temperature threshold t3, and a dry-burning gradient ΔT; define the temperature less than t1 as the first temperature threshold range; define the temperature greater than or equal to t1 and less than t2 as the second temperature threshold range; define the temperature greater than or equal to t2 and less than t3 as the third temperature threshold range; define the temperature greater than or equal to t3 as the fourth temperature threshold range. Initially, the device enters the temperature tracking mode in a low-frequency sampling state, and let the thermal image collected at this time be img, and the highest temperature value of the pixels in the thermal image img is t max = max(img); If t max < t1, continuously execute the temperature tracking mode of step 1 in the low-frequency sampling state; when t1 ≤ t max , enter step 2.
3. The thermal imaging anti-dry-burning method based on state transition according to claim 2, characterized in that: Step 2 specifically includes the following steps: When t1≤t max At that time, the system switches from low-frequency sampling to high-frequency sampling and initializes the number of valid pixels C. effect =0, initialize the valid pixel image img effect If the value is empty, initialize the first counter C1 = 20; Statistical analysis of the highest temperature value t of pixels in a thermal image (img). max , and the number of effective pixels C in the thermal image img that exceed the first temperature threshold t1; Compare C and C effect If C>C effect Then C effect =C, making img effect =img; if C≤C effect , do not make C effect and img effect Update; If the highest temperature value t in the thermal image collected in the high-frequency sampling state max <t1, then for each collection of a thermal image, make C1 = C1 - 1 until C1 = 0, return to the temperature tracking mode of step 1, and switch to the low-frequency sampling state; If the highest temperature value t1 in the thermal image collected in the high-frequency sampling state ≤ t max <t2, then keep C1 = 20 and continue with the temperature tracking mode in step 2 under the high-frequency sampling state; If the highest temperature value t2 in the thermal image collected in the high-frequency sampling state ≤ t max <t3, set C1 = 20 and enter the gradient detection mode of step 3; If the highest temperature value t3 ≤ t in the thermal image acquired under high-frequency sampling conditions max Make C1 = 20, C dog =20, C dog For the watchdog counter, proceed to step 4, watchdog detection mode.
4. The thermal imaging anti-dry-burning method based on state transition according to claim 3, characterized in that: Step 3 specifically includes the following steps: When t2 ≤ t max <When t < t3, the temperature tracking mode in the high-frequency sampling state is changed to the gradient detection mode. The gradient detection mode takes the thermal images collected in N seconds as a group, and constructs a sequence T of the highest temperature values t max <in the thermal images. Let n <represent the highest temperature value in the i-th thermal image in this group, where i ∈ N; <denote the highest temperature value in the i-th thermal image of this group, with i ∈ N; In the first N seconds, sequence T n When the number of samples in sequence T is less than N, thermal images are continuously acquired in a high-frequency sampling state until a full set of thermal images is acquired; when sequence T... n When the number of elements in a sequence equals N, calculate the average value M of the sequence. n When sequence T n When the number of elements in sequence T is greater than N, the sequence T is replaced sequentially. n The highest temperature value in the sequence is used to calculate the average value M of the group sequence. n+1 Simultaneously calculate the average value M of the group sequence. n+1 Compared with the mean M of the previous group sequence n The difference ΔM = M n+1 -M n ; If ΔM < 0, then thermal images are continuously acquired according to the high-frequency sampling state, and sequence T is replaced sequentially. n The highest temperature value in the sequence; if ΔM≥0, then for sequence T n Perform order-preserving transformation; Order-preserving transformation is a transformation of a sequence Abstracted as X = {x i ,x i+1 ,...,x i+14 }, check the monotonicity of sequence X; If the sequence X does not satisfy monotonically increasing, then when the sequence X has x at the k-th position k >x k+1 , and x k+1 >x m , when i ≤ k < i + 14, m ∈ {i, i + 1,..., i + 12} and m < k, construct a new sequence: Let X' = X. If X' still does not satisfy the monotonically increasing condition, repeat the order-preserving transformation. If X or X' satisfies the monotonically increasing condition, calculate the temperature increase Δt = x of the sequence X or X'. i+14 -x i If Δt > ΔT, then it is determined that the dry burning state is imminent and a warning is issued; otherwise, the sequence X is cleared and the process returns to step 2.
5. The thermal imaging anti-dry-burning method based on state transition according to claim 3, characterized in that: Step 4 specifically includes the following steps: When t3≤t max At that time, the temperature tracking mode changes from high-frequency sampling to watchdog detection mode. In watchdog detection mode, the newly acquired thermal images img and img are compared. effect Statistics in img effect The value in the image represents the fourth temperature threshold range, but in the image itself it represents the number of pixels within the first temperature threshold range (count). effect : count effect =count[{(i,j)|img(i,j)<t1 and img effect (i,j)≥t3}] If count effect ≤5, making C dog =C dog -1, and return to step 2, but keep the updated C dog If count effect >5, then C dog =30, and return to step 2, but keep the updated C. dog Update C dog Afterwards, if C dog If the value is 0, it is determined to be a dry-burning state, a warning is issued, and the valve is shut off.
6. A state-transfer-based thermal imaging anti-dry-burning device, applied to the method of any one of claims 1-5, characterized in that: Includes monitoring module, control module, thermocouple valve, power supply unit, voltage regulator circuit, and communication module; The monitoring module is a miniature thermal imaging camera, used to capture thermal images of the stove and upload the thermal images to the control module via a communication module; The control module is used to implement the method according to any one of claims 1-5; The power supply unit supplies power to the monitoring module, control module, thermocouple valve, and communication module through a voltage regulator circuit. When the control module determines that the stove is in a dry-burning state, the thermocouple valve will shut off.