JP2006189403A - Frosting prediction device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enhance hit precision in frosting prediction, as to a frosting prediction device. <P>SOLUTION: The present invention adopts a means of predicting frosting in a frosting prediction objective area, when satisfying all the frosting determination conditions as to one or both of an ambient temperature going-down degree and a radiation balance, in addition to the respective frosting determination conditions as to an ambient temperature, a dew point temperature, an average wind velocity and a relative humidity, out of a meteorological forecasting data in the frosting prediction objective area. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a frost prediction apparatus.

  In cold regions, the damage to crops caused by frost is enormous, which is a serious problem. Therefore, prediction of frost in the agricultural field is very important in preventing damage caused by frost. The conventional frost forecasting method uses the frost judgment conditions regarding the temperature, the estimated surface temperature (temperature-3.5 ° C), the dew point temperature, the average wind speed, and the relative humidity in the weather forecast data of the frost forecast target area. I was doing. For example, according to the frost determination conditions obtained using Matsuo Village in Iwate Prefecture as the test area, temperature <3.5 ° C, temperature -3.5 ° C <dew point temperature, average wind speed <1.5 m / s and relative humidity> It was predicted that frost would form when all four conditions of 95% were satisfied.

  However, when frost prediction is performed according to the above four frost determination conditions, there is a problem that the accuracy of prediction is low.

 This invention is made | formed in view of such a situation, and aims at the improvement of the accuracy of the frost prediction.

In order to solve the above-mentioned problems, in the present invention, as the first solving means relating to the frost prediction device, among the frost prediction target area, the frost determination conditions relating to the temperature, dew point temperature, average wind speed and relative humidity. In addition, when all the frost determination conditions regarding a temperature fall degree and a radiation balance are satisfied, a means of predicting frost in the frost prediction target area is adopted.
According to the present invention, the frost prediction is performed by adding the frost determination conditions related to the temperature drop degree and the radiation balance to the frost determination conditions related to the conventional temperature, dew point temperature, average wind speed and relative humidity. Can be improved.

Further, as a second solving means related to the frost prediction device, the temperature Ta, the relative humidity H, the average wind speed W, the ground surface temperature Tg and the radiation balance Ra, which are the weather forecast data of the frost prediction target area, and the respective weather forecast data With respect to the following frost judgment formula (1) for obtaining the frost judgment value Y from the weighting coefficients K1 to K5 and the constant C, the weighting coefficient K1 is calculated by calculating a plurality of local weather data in the frost prediction target area based on the discriminant analysis method. The means of specifying ~ K5 and constant C and predicting frost in the frost prediction target area is adopted.
According to this invention, since frost prediction is performed by the frost determination formula obtained by discriminating and analyzing a plurality of past local weather data, it is possible to perform statistically high-precision prediction. Furthermore, it is possible to update the frost determination formula by continuously collecting the local weather data and performing the discriminant analysis, and it is possible to perform the frost prediction by the frost determination formula suitable for the weather conditions of the frost prediction target area. .

  Y = K1, Ta + K2, H + K3, W + K4, Tg + K5, Ra + C (1)

Further, as a third solving means related to the frost prediction device, the frost prediction process in the first solving means and the frost prediction process in the second solving means are set to the accuracy of the frost prediction with respect to the actual presence or absence of frost. A method of switching based on the above is adopted.
According to this invention, by selecting the frost prediction process with the higher predictive predictive ratio when the frost is predicted using the frost determination condition and when the frost is predicted using the frost determination formula based on the discriminant analysis, The frost prediction suitable for the frost prediction target area can be performed, and the predictive medium accuracy can be increased.

Further, as a fourth solving means related to the frost prediction apparatus, in any one of the first to third solving means, an adhesion state of frost to a predetermined observation object is acquired as image data, and the image data is binarized. The means of determining the presence / absence of actual frost by performing the conversion process is adopted.
Conventionally, the presence or absence of actual frost has been determined by human eyes, but according to the present invention, the determination of the presence or absence of frost can be automated, which is effective in reducing labor costs.

Further, as a fifth solving means related to the frost prediction apparatus, a means of adding a desiccant to the observation object in the fourth solving means is adopted.
In some cases, it is difficult to determine whether the surface condition of the observation object is frost or condensation. According to this invention, condensation can be prevented by adding a desiccant to the observation object, and misjudgment of frost or condensation can be prevented.

Further, as a sixth solving means related to the frost prediction apparatus, in the fourth or fifth solving means, a binarization process is performed after removing the halation contained in the image data of the observation object. adopt.
If halation is included in the image data, it may be difficult to determine whether the data is a frosted portion or a halation portion by subsequent binarization processing. According to the present invention, the binarization process is performed after removing the halation contained in the image data.
Only image data of a portion with frost can be extracted, and misjudgment with halation can be prevented.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a schematic diagram of a frost prediction system using the frost prediction apparatus of the present invention. In this figure, reference numeral 1 is a frost prediction device, 2 is a network, 3 is a weather forecast server, 4 is a client computer, and 5 is a local weather observation device.

 The frost prediction apparatus 1 includes a frost prediction calculation unit 1a, a communication unit 1b, and a storage unit 1c. The frost prediction calculation unit 1a acquires weather forecast data of the frost prediction target area from the communication unit 1b, performs frost prediction based on the frost determination condition or the frost determination formula, and sends the frost prediction result to the communication unit 1b. . Moreover, the local frost prediction calculation part 1a is connected with the local weather observation apparatus 5, and the frost prediction calculation part 1a acquires local weather data from the local weather observation apparatus 5, and memorize | stores it in the memory | storage part 1c. The frost prediction operation of the frost prediction calculation unit 1a will be described in detail later.

 The communication unit 1 b is connected to the network 2, and is connected to the weather forecast server 3 and the client computer 4 via the network 2. The communication unit 1b acquires the weather forecast data of the frost prediction target area from the weather forecast server 3 via the network 2, and sends the frost forecast result obtained by the frost forecast calculation unit 1a to the client computer 4 via the network 2. Is.

 The weather forecast server 3 is installed in, for example, a private weather forecaster and provides weather forecast data. The client computer 4 is installed in a farmer or an agriculture-related facility, and informs an agricultural person about the frost prediction result.

  The local meteorological observation device 5 includes an air temperature measuring unit 5a, a relative humidity measuring unit 5b, an average wind speed measuring unit 5c, a surface temperature measuring unit 5d, a radiation balance measuring unit 5e, a dew point temperature measuring unit 5f, a frost observation device 5g, and local meteorological data. It consists of a transmitter 5h.

The local meteorological observation device 5 is installed in the frost prediction target area, and whether or not the frost actually falls
Regardless, local weather data is continuously collected. That is, the temperature, the relative humidity, the average wind speed, the surface temperature, and the radiation are measured by the temperature measuring unit 5a, the relative humidity measuring unit 5b, the average wind speed measuring unit 5c, the surface temperature measuring unit 5d, the radiation balance measuring unit 5e, and the dew point temperature measuring unit 5f. The balance and dew point temperature are measured, and whether or not frost has actually been observed is observed by the frost observation device 5g. Conventionally, whether or not frost has actually been confirmed is confirmed by human eyes, but in this embodiment, automatic observation is enabled by the frost observation device 5g. The frost observation device 5g will be described later in detail. These measurement results and observation results are sent as local meteorological data from the local meteorological data transmitting unit 5h to the frost forecast calculating unit 1a of the frost forecasting apparatus 1.

Next, the frost prediction operation | movement of the frost prediction apparatus 1a is demonstrated in detail.
FIG. 2 is a flowchart showing the frost prediction operation. As shown in this figure, the frost prediction calculation unit 1a first selects whether to perform frost prediction using frost prediction processing of a frost determination condition or a frost determination formula as an initial setting. In the present embodiment, it is assumed that frost prediction is first performed using frost determination conditions (step S1).

  When the frost prediction process is determined, the frost prediction calculation unit 1a acquires the weather forecast data of the frost prediction target area from the weather forecast server 3 via the communication unit 1b (step S2), and the weather forecast data and the frost determination. The frost prediction is performed based on the conditions (step S3). Step S3 will be described in detail below.

 Conventionally, the frost determination conditions used in step S3 were predicted to generate frost when all the frost determination conditions regarding the air temperature, dew point temperature, average wind speed, and relative humidity were satisfied. Specifically, these frost detection conditions are as follows: air temperature <3.5 ° C., air temperature −3.5 ° C <dew point temperature, average wind speed <1.5 m / s, which was obtained in Matsuo Village, Iwate Prefecture. And relative humidity> 95%. When frost prediction is performed under these frost determination conditions, a result with a predictive predictive ratio of 83.1% is obtained, but it is necessary to further improve predictive predictive accuracy in order to prevent frost damage to the maximum extent.

 Therefore, in addition to the conventional frost determination conditions, the frost determination conditions used in the present embodiment satisfy all the frost determination conditions related to both the temperature drop (temperature of the frost prediction target time—the temperature before 1 hour) and the radiation balance. In some cases, it is predicted that frost will form. Specifically, the frost determination conditions related to the temperature and relative humidity are also corrected, and the temperature <2 ° C, temperature −3.5 ° C <dew point temperature, average wind speed <1.5 m / s and relative humidity> It is predicted that frost will form when all 85% are satisfied and in addition both the temperature drop <−0.2 ° C. and the radiation balance <−0.2 MJ. According to such frost determination conditions, it is possible to obtain a predictive medium accuracy of a predictive medium rate of 92.2%.

  As described above, in step S3, the temperature, dew point temperature, average wind speed, relative humidity, temperature drop degree (that is, the temperature one hour ago) and radiation balance in the weather forecast data are applied to the above six frost determination conditions. Perform the action. The frost prediction calculation unit 1a determines that frost will be generated if the weather forecast data satisfies all the frost determination conditions, and determines that frost will not be generated if all are not satisfied (step S4). And if it determines with frost falling, an alarm will be sent to the client computer 4 via the communication part 1b, and the prediction result that frost will fall is notified to the agriculture official (step S5).

  Next, the frost prediction calculation part 1a acquires local weather data from the local weather observation apparatus 5 at the date and time when the frost prediction is made, regardless of the prediction result (step S6). In the local meteorological data, frost judgment is performed by applying the frost judgment conditions to the temperature (including the temperature one hour ago), dew point temperature, average wind speed, relative humidity, and radiation balance. By comparing with the observation result of frost, it is stored in the storage unit 1c as a predictive intermediate result of the day (step S7).

  Next, the frost prediction calculation part 1a judges whether it is a predictive middle rate calculation time (step S7a), and if it is not the time yet, it will return to step S2. As described above, the operations from S2 to S7a are repeated for a certain period (until the predictive predictive ratio calculation time), and the local weather data and the predictive intermediate result within that period are sequentially stored for each date and time in the storage unit 1c. Is done. When the predictive predictive rate calculation time comes, the predictive predictive value when the frost prediction calculating unit 1a predicts frost using the frost determination condition in that period based on the predictive predictive result stored in the storage unit 1c. A is calculated (step S8).

  At this time, the frost prediction calculation unit 1a calculates a predictive medium rate A, and also calculates a predictive medium rate B when another frost prediction process, which is a frost prediction formula, is used. The procedure will be described below.

  When the frost determination condition is selected as the initial setting of step S1, the operations from step S10 to S14 are not actually performed, and the predictive probability B is calculated by the operations of steps S9 and S15 to S18. However, in order to explain how the frost prediction is performed using the frost determination formula, description will be made in order from step S9.

 The frost determination formula is a formula for determining the frost determination value Y using the temperature Ta, the relative humidity H, the average wind speed W, the ground surface temperature Tg and the radiation balance Ra as variables as in the following formula (1). If it is 0, it determines with frost falling. The weighting coefficients K1 to K5 and the constant C of each variable are specified by performing arithmetic processing on a plurality of past local weather data in the frost prediction target area based on the discriminant analysis method (step S9). Further, in this step S9, the frost determination formula is updated at any time by calculating the local weather data stored in the storage unit 1c based on the discriminant analysis method at regular intervals in step S17 described later. Frost prediction based on the local weather data can be performed, and predictive medium accuracy can be improved.

    Y = K1, Ta + K2, H + K3, W + K4, Tg + K5, Ra + C (1)

  In the present embodiment, arithmetic processing based on the discriminant analysis method is performed on past local weather data, and weighting coefficients K1 to K5 and a constant C are calculated to obtain Expression (2). According to this formula (2), it is possible to increase the predictive median rate of 85.1% to the predictive median rate of 83.1% when the frost prediction is performed under the conventional frost determination conditions.

  Y = 1.5462Ta-0.0341H-0.0291W-0.7467Tg + 4.2318Ra + 7.3716 (2)

  And the frost prediction calculation part 1a will acquire the weather forecast data of a frost prediction target area from the weather forecast server 3 via the communication part 1b, if a frost determination formula is determined as mentioned above (step S10). In the weather forecast data, if the temperature is higher than 8 ° C, it can be determined that frost does not fall. Therefore, in step 11, it is determined whether the temperature is 8 ° C or lower. If the temperature is 8 ° C. or lower, the operation of the next step S12 is performed. Here, when the explanation is continued assuming that the air temperature is 8 ° C. or less, the frost prediction calculating unit 1a substitutes the weather forecast data, that is, the air temperature, the relative humidity, the average wind speed, the ground surface temperature, and the radiation balance into the equation (2). A frost determination value Y is obtained (step S12). When this frost determination value Y is smaller than zero, it determines with frost falling (step S13), sends an alarm to the client computer 4 via the communication part 1b, and notifies the agricultural person concerned about the prediction result that frost falls (step) S14).

 Next, the frost prediction calculation part 1a acquires local weather data from the local meteorological observation device 5 on the date and time when the frost prediction is made regardless of the prediction result (step S15). Then, the frost determination is performed based on the frost determination formula (2) and the local meteorological data, and is stored in the storage unit 1c as a predictive intermediate result of the day by comparing with the frost observation result of whether or not the frost actually falls. (Step S16).

  Next, the frost prediction calculation part 1a determines whether it is the update time of a frost determination formula (step S17), and if it is an update time, it will be discriminant-analyzed based on the local weather data memorize | stored in the memory | storage part 1c in the period until now. A new frost determination formula is obtained by the method (step 9), and frost prediction is performed from the next time using the updated frost determination formula.

  Next, the frost prediction calculation part 1a judges whether it is a predictive middle rate calculation time (step S17a), and if it is not the time yet, it will return to step S10. As described above, when the frost determination condition is selected as the initial setting in step S1, the operations in steps S10 to S14 are not actually performed, and after the frost determination formula is specified in step S9, step S15 and step S15 are performed. The operation of step S17a is repeated for a certain period, and the local weather data and the predictive intermediate result within the period are sequentially stored for each date and time in the storage unit 1c.

 And if a fixed period passes, the frost prediction calculating part 1a will calculate the predictive medium rate B at the time of frost prediction using the frost judging formula in the period based on the predictive intermediate result memorize | stored in the memory | storage part 1c. (Step S18).

  The frost prediction calculation unit 1a compares the predictive median A and the predictive median B obtained as described above, and switches to the frost prediction process with the higher predictive median to perform the next frost prediction. Perform (step S19). Here, if the predictive predictive value A and the predictive predictive value B are the same value, the frost prediction process is not switched. For example, if the predictive predictive value B is high, the operation is performed from step S9 next, and the operations of steps S9 to S17a are repeated for a certain period. In that case, the operation from step S2 to step S5 is not performed, and the operation from step S6 to S7a is repeated. Then, after a certain period of time, the predictive predictive value A and the predictive predictive value B are calculated again, and the selection of the frost prediction process is performed.

  When the frost determination formula is selected as the initial setting of Step S1, the operation is performed from Step S9, and the operations of Step S9 to Step S17a are repeated for a certain period. In that case, the operation from step S2 to step S5 is not performed, and the operation from step S6 to S7a is repeated. When a certain period of time elapses, the predictive predictive value A and the predictive predictive value B are calculated, and the frost prediction process is selected.

 As described above, according to the present embodiment, by selecting a frost prediction process having a high predictive predictability for each predetermined period, it is possible to perform frost prediction suitable for the topography and weather conditions in the frost prediction target area, and predictive Medium accuracy can be increased.

  Moreover, in this embodiment, in order to fully automate the above frost prediction system, the frost observation instrument 5g is used. Conventionally, it has been confirmed by human eyes whether or not frost has formed. However, since it is necessary to check frost from the night when the temperature is low to the early morning, it is a heavy burden on the operator. From this point of view, the use of the frost observing instrument 5g has an advantage of not only automation of the system but also labor cost reduction. Hereinafter, the frost observation instrument 5g will be described.

FIG. 3 is a schematic diagram showing the configuration of the frost observation instrument 5g. Reference numeral 20 is an infrared camera, 21 is a glass petri dish (observation object) for attaching frost, and 22 is a glass petri dish base. The infrared camera 20 is installed in the observation hut 23 so that the surface state of the glass petri dish 21 can be photographed, and is connected to the local weather data transmission unit 5 h by a connection cable 24. The image data obtained by the infrared camera 20 is sent from the local weather data transmission unit 5h to the frost prediction calculation unit 1a, and it is determined whether or not frost has been formed by image processing.

In addition, about the surface state of the glass petri dish 21, it may be difficult to judge whether frost adheres or dew condensation. Therefore, in this embodiment, silica gel is put in the glass petri dish 21 as a desiccant. Thereby, condensation of the glass petri dish 21 can be prevented, and erroneous determination of frost or condensation can be prevented.

 Further, the infrared projection of the infrared camera 20 is reflected on the glass petri dish 21 to cause halation, which may hinder observation of frost attached to the glass petri dish 21. In order to reduce the influence of this halation, in this embodiment, the installation location of the infrared camera 20 is set as follows: the installation angle θ of the infrared camera 20 is 45 °, the distance d between the infrared camera 20 and the glass petri dish 21 is 140 cm, as shown in FIG. The installation height h1 of the infrared camera 20 was 120 cm, and the height h2 of the glass petri dish base 22 was 22 cm.

Hereinafter, the image processing operation of the frost prediction calculation unit 1a as described above will be described with reference to the flowchart of FIG.
The frost prediction calculation part 1a acquires the image data of the surface state of the glass petri dish 21 photographed by the infrared camera 20 from the local weather data transmission part 5h (step S20). Here, since the infrared projection of the infrared camera 20 is reflected on the glass petri dish 21 to cause halation, the image data also includes halation, and when performing the following image processing, it is erroneously determined that frost is attached. There is a risk. Therefore, the halation portion is cut out from the image data (step S20a). Next, a binarization process is performed in order to distinguish a portion with frost and a portion without frost in the image data (step S21). An area ratio between the frosted portion and the frostless portion is calculated from the binarized image data (step S22). Next, the above-mentioned area ratio is compared with a preset set value (step S23), and if the area ratio> set value, it is determined that frost has fallen (step S24). Otherwise, no frost has fallen. (Step S25). These determination results are stored in the storage unit 1c as frost observation results (step S26) and used when calculating the predictive predictive value.

 As described above, in the present embodiment, the influence of halation is prevented by the installation position of the infrared camera 20 and image processing, but an IR filter may be provided in the infrared camera 20 as another method.

  As described above, according to the present embodiment, not only an improvement in frost predictive medium accuracy but also an automation of the frost prediction system can be realized.

In addition, this invention is not limited to the said embodiment, For example, the following modifications can be considered.
(1) As frost determination conditions, the temperature <2 ° C, the temperature -3.5 ° C <dew point temperature, the average wind speed <1.5 m / s and the relative humidity> 85% are all satisfied, and the temperature drop < Although it was predicted that frost would be generated when both −0.2 ° C and radiation balance <−0.2 MJ were satisfied, it is not limited to the above values, and the topography and weather conditions of the frost prediction target area In consideration of the various characteristics, it may be changed to a numerical value that can predict frost most accurately for each region.

(2) When determining the frost determination formula (1) by discriminating and analyzing a plurality of local weather data, F ratio corresponding to each variable, temperature Ta, relative humidity H, average wind speed W, surface temperature Tg, and radiation balance Ra May be calculated, and only the variable having the F ratio of 2.0 or more may be used in the frost determination formula (1). Here, the F ratio represents how significant the variable has in determining the frost determination value Y, and the larger the F ratio, the more significant the variable. Therefore, Equation (1) can be simplified by not using a variable having a small F ratio.

(3) If the predictive predictive value A and the predictive predictive value B are the same value in step S19 of FIG. 2, the frost prediction process is not switched. However, it is also possible to calculate the rate of frost formation) and switch to the frost prediction process with the lower miss rate. In this case, if the miss rate is the same value, the frost prediction process is not switched.

(4) In the above embodiment, when the frost prediction calculation unit 1a predicts that frost will be generated, it merely outputs an alarm to the client computer 4 via the network 2. However, the present invention is not limited to this. A frost countermeasure device using an equation or combustion type is provided, the frost countermeasure device is connected to the frost prediction calculation unit 1a, and the frost prediction calculation unit 1a outputs an alarm and simultaneously outputs a start signal of the frost countermeasure device. By doing so, you may be made to automatically take measures against frost on site. By doing in this way, unmanned and automated can be realized consistently from frost prediction to on-site measures. Further, the frost countermeasure device may be connected to the frost prediction device 1 via the network 2.

It is a lineblock diagram of the frost prediction system concerning one embodiment of the present invention. It is a frost prediction operation | movement flowchart figure of the frost prediction apparatus in this embodiment. It is a block diagram of the frost observation instrument in this embodiment. It is an image processing operation | movement flowchart figure of the frost prediction apparatus in this embodiment.

Explanation of symbols

 DESCRIPTION OF SYMBOLS 1 ... Frost prediction apparatus, 2 ... Network, 3 ... Weather forecast server, 4 ... Client computer, 5 ... Local weather observation apparatus

Claims (6)

  In the weather forecast data of the frost forecast target area, in addition to the frost judgment conditions related to temperature, dew point temperature, average wind speed and relative humidity, in addition to satisfying all the frost judgment conditions related to temperature drop and radiation balance, the frost forecast A frost prediction apparatus for predicting frost in a target area. From the temperature Ta, relative humidity H, average wind speed W, surface temperature Tg and radiation balance Ra, which is the weather forecast data of the frost forecast target area, and the weighting coefficients K1 to K5 and the constant C relating to each weather forecast data, the frost judgment value Y is obtained. About the following frost judgment formula to be obtained, a plurality of local meteorological data in the frost prediction target area is processed based on the discriminant analysis method to identify the weighting coefficients K1 to K5 and the constant C to predict frost in the frost prediction target area. The frost prediction apparatus characterized by performing.
Y = K1, Ta + K2, H + K3, W + K4, Tg + K5, Ra + C (1)   The frost prediction apparatus according to claim 1, wherein the frost prediction process in claim 1 and the frost prediction process in claim 2 are switched based on an accuracy of frost prediction with respect to the presence or absence of actual frost.   The frost adhesion state to a predetermined observation object is acquired as image data, and the presence or absence of actual frost is determined by binarizing the image data. The frost prediction apparatus of description.   The frost prediction apparatus according to claim 4, wherein a desiccant is added to the observation object. The frost prediction apparatus according to claim 4 or 5, wherein binarization processing is performed after removing halation contained in image data of the observation object.

Images