Wastewater treatment management system, wastewater treatment management method

Abstract

To provide a wastewater treatment management system capable of greatly reducing the burden of a manager and optimizing the charge amount of a flocculant for generating floc.SOLUTION: A wastewater treatment management system manages a state of treated water W1 obtained by reacting wastewater with a flocculant to generate floc 31. The wastewater treatment management system includes detection model construction means, floc detection means, and prediction model construction means. The detection model construction means constructs a detection model for detecting the floc 31 from an image 41 of the treated water W1. The flock detection means detects the floc 31 in an analysis region 43 set in the image 41, by performing image processing using the detection model in the analysis region 43. The prediction model construction means constructs a prediction model for predicting a size of the floc 31 after a predetermined time, by accumulating data of sizes of the detected floc 31.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a wastewater treatment management system and a wastewater treatment management method for managing the state of treated water obtained by reacting wastewater with a flocculant to generate flocs. [Background technology]

[0002] Conventionally, wastewater treatment facilities have been known that remove harmful substances (e.g., fluorine and heavy metals such as zinc) contained in wastewater discharged from painting facilities by precipitating them as flocs. In such wastewater treatment facilities, an on-site manager visually inspects the condition and composition of treated water obtained by reacting a coagulant with the wastewater to generate flocs and by managing the pH. Furthermore, various technologies for managing the condition of treated water have been proposed (see, for example, Patent Documents 1 and 2). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 9-281100 [Patent Document 2] Japanese Patent Application Laid-Open No. 2005-270689 Summary of the Invention [Problem to be solved by the invention]

[0004] Meanwhile, the manager inspects the condition and composition of the treated water at regular intervals (for example, every hour). However, the large number of inspection items places a heavy burden on the manager, and there is a risk of overlooking an inspection. Furthermore, the inspection area is large and has multiple floors, and the manager must look into the treated water to check its condition, which also increases the manager's burden. Furthermore, the manager determines the amount of flocculant to be added to generate flocs and the operation of the pump to replenish the flocculant, based on the condition of the treated water. However, these decisions are largely based on the manager's experience, so there is a risk of quality variations depending on the manager.

[0005] The present invention has been made in view of the above-mentioned problems, and an object of the present invention is to provide a wastewater treatment management system and a wastewater treatment management method that can significantly reduce the burden on an administrator and optimize the amount of coagulant to be added to generate flocs. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems, the invention described in claim 1 is a wastewater treatment management system that manages the state of treated water obtained by reacting a coagulant with wastewater to generate flocs in wastewater treatment equipment that removes harmful substances contained in the wastewater discharged from a painting equipment by precipitating them as flocs, the system comprising: a detection model construction means that constructs a detection model to detect the flocs from an image of the treated water; a floc detection means that detects the flocs in an arbitrary analysis area preset in the image by performing image processing using the detection model in the analysis area; and a prediction model construction means that constructs a prediction model that predicts the size of the flocs after a predetermined time by accumulating data on the sizes of the detected flocs.

[0007] In the invention described in claim 1, the detection model detects flocs within an analysis area set in an image of treated water, and the prediction model predicts the size of the flocs after a predetermined time. This automates sensory evaluation (specifically, evaluation of the coarsening state of flocs), which was previously performed visually by the manager, and eliminates the need for the manager to look into the treated water, significantly reducing the burden on the manager. In addition, the amount of coagulant to be added can be optimized based on the floc size predicted by the prediction model, making it economical.

[0008] The invention described in claim 2 is characterized in that, in claim 1, the prediction model creation means further comprises a digitizing means for digitizing the size of the flocs detected by the floc detection means, and the prediction model creation means accumulates data on the floc sizes digitized by the digitizing means.

[0009] In the invention described in claim 2, the predictive model construction means accumulates data on the floc sizes quantified by the quantification means, thereby making it possible to obtain the floc sizes after a predetermined time with high accuracy.

[0010] The invention described in claim 3 is based on claim 2 and further comprises a data management means for managing data pairs each including the number of flocs detected in the analysis area by the floc detection means and the mode of the length of the long side of a bounding box surrounding the detected flocs, and the digitization means digitizes the size of the flocs based on the data managed by the data management means.

[0011] In the invention described in claim 3, the quantification means quantifies floc size based on data managed by the data management means, i.e., data including the number of detected flocs and the most frequent value of the length of the long side of the bounding box surrounding the floc. Therefore, by checking the number of detected flocs within the analysis area included in the data, it is possible to estimate the size of each floc within the analysis area. Furthermore, by checking the most frequent value of the length of the long side of the bounding box included in the data, it is possible to estimate the maximum length of the floc in a plan view. Therefore, by accumulating the floc size data quantified by the quantification means in a prediction model, it is possible to obtain a highly accurate prediction value of the floc size after a predetermined time.

[0012] The invention described in claim 4 is based on claim 1 and further comprises an imaging device that is installed in a treatment tank that stores the treated water and captures an image of the treated water to obtain the image.

[0013] In the invention described in claim 4, flocs contained in the treated water stored in the treatment tank can be reliably detected based on an image taken from close up, and the size of the flocs after a predetermined time can be accurately predicted.

[0014] The invention described in claim 5 is based on claim 1 and further comprises an input amount calculation means for calculating the input amount of the flocculant that will result in the size of the flocs after a predetermined time by performing an optimization calculation using the prediction model.

[0015] In the invention described in claim 5, the dosage calculation means automatically calculates the dosage of flocculant that will result in the floc size after a predetermined time based on the floc size predicted by the prediction model. Therefore, the dosage of flocculant can be obtained more easily and accurately than when the manager decides the dosage amount.

[0016] The gist of the invention described in claim 6 is that, in claim 1, when the size of the flocs detected by the floc detection means differs from the actual size of the flocs contained in the treated water, the prediction model is automatically updated by correcting the data on the size of the flocs.

[0017] According to the invention of claim 6, even if the floc size detected by the floc detection means differs from the actual floc size, the prediction model can be quickly and automatically updated, thereby enabling accurate floc size predictions to be quickly obtained.

[0018] The invention described in claim 7 is a method for managing the state of treated water obtained by reacting a coagulant with wastewater discharged from a painting facility to generate flocs and remove harmful substances contained in the wastewater by precipitating the flocs, the method comprising: a detection model construction step of constructing a detection model to detect the flocs from an image of the treated water; a floc detection step of detecting the flocs in an arbitrary analysis region preset in the image by performing image processing using the detection model; and a prediction model construction step of constructing a prediction model to predict the size of the flocs after a predetermined time by accumulating data on the sizes of the detected flocs.

[0019] In the invention described in claim 7, the detection model constructed in the detection model construction step detects flocs within an analysis area set in an image of the treated water, and the prediction model constructed in the prediction model construction step predicts the size of the flocs after a predetermined time. This automates sensory evaluation (specifically, evaluation of the coarsening state of flocs), which was previously performed visually by the manager, and eliminates the need for the manager to look into the treated water, significantly reducing the manager's burden. Furthermore, the amount of coagulant to be added can be optimized based on the floc size predicted by the prediction model, which is economical. [Effects of the Invention]

[0020] As described above in detail, according to the inventions set forth in claims 1 to 7, the burden on the manager can be significantly reduced, and the amount of flocculant to be added for generating flocs can be optimized. [Brief explanation of the drawings]

[0021] [Figure 1] 1 is a schematic configuration diagram showing a wastewater treatment management system according to an embodiment of the present invention. [Figure 2] (a) is a photograph showing an image of the treated water, and (b) is an enlarged view showing the analysis area. [Figure 3] (a) is a photograph showing an image of treated water when 97 flocs were detected, (b) is an enlarged view showing the analysis area, and (c) is an enlarged view showing the bounding box. [Figure 4] (a) is a photograph showing an image of treated water when 37 flocs were detected, (b) is an enlarged view showing the analysis area, and (c) is an enlarged view showing the bounding box. [Figure 5] 10 is a table showing data paired with the number of detected flocs and the mode of the length of the long side of the bounding box. [Figure 6] FIG. 4 is a schematic diagram showing a display screen of a display on which a management form is displayed. [Figure 7] FIG. 10 is a schematic diagram showing a display screen on which the status and schedule of the control panel are displayed. [Figure 8] Schematic diagram showing a display screen on which a pH value is displayed. [Figure 9] 1 is a schematic diagram illustrating a display screen displaying a gallery of images. DETAILED DESCRIPTION OF THE INVENTION

[0022] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will now be described in detail with reference to the accompanying drawings.

[0023] As shown in Fig. 1, the wastewater treatment management system 1 of this embodiment is a system that manages the state of treated water W1 in a wastewater treatment facility 10. The wastewater treatment facility 10 is a facility that removes harmful substances (in this embodiment, heavy metals such as zinc and fluorine) contained in wastewater W2 discharged from a coating facility 2 by precipitating them as flocs 31 (see Fig. 2). The treated water W1 is obtained by reacting the wastewater W2 with coagulants F1 and F2 to generate the flocs 31.

[0024] The wastewater treatment facility 10 also includes a raw water tank 11, a first reaction tank 12, a second reaction tank 13, a coagulation tank 14 (treatment tank), a pressurized flotation tank 15, a sludge tank 16, a dehydrator 17, a pH adjustment tank 18, and a discharge tank 19. The raw water tank 11 stores wastewater W2 discharged from the coating facility 2, and supplies the stored wastewater W2 to the first reaction tank 12.

[0025] The first reaction tank 12 stores wastewater W2 supplied from the raw water tank 11. A positively charged flocculant F1 is supplied from the first tank 21 to the wastewater W2 stored in the first reaction tank 12. This neutralizes the charge and causes zinc in the wastewater W2 to flocculate. In this embodiment, polyferric sulfate, an iron-based inorganic polymer flocculant, is used as the flocculant F1. By adding slaked lime to the wastewater W2 stored in the first reaction tank 12, fluoride ions in the wastewater W2 react with calcium ions (slaked lime) to precipitate calcium fluoride, which then precipitates. The first reaction tank 12 then supplies the wastewater W2, in which zinc has flocculated and calcium fluoride has precipitated, to the second reaction tank 13.

[0026] As shown in FIG. 1, the second reaction tank 13 stores wastewater W2 supplied from the first reaction tank 12. A flocculant F2 is supplied from the second tank 22 to the wastewater W2 stored in the second reaction tank 13. This increases the pH value of the wastewater W2, increasing the hydroxide ion concentration. As a result, zinc hydroxide, an insoluble hydroxide, precipitates to form flocs 31 (see FIG. 2). In this embodiment, caustic soda is used as the flocculant F2. By adding aluminum sulfate to the wastewater W2 stored in the second reaction tank 13, fluorine in the wastewater W2 is adsorbed by the aluminum contained in the aluminum sulfate, resulting in fluorine precipitation. The second reaction tank 13 then supplies treated water W1 obtained by generating flocs 31 to the coagulation tank 14.

[0027] The coagulation tank 14 stores the treated water W1 supplied from the second reaction tank 13. A polymer coagulant F3 is supplied from a polymer coagulant dissolving tank 23 to the treated water W1 stored in the coagulation tank 14. As a result, a large number of flocs 31 contained in the treated water W1 are entangled by the polymer coagulant F3 and coarsened. The coagulation tank 14 then supplies the treated water W1 with the coarsened flocs 31 to the pressurized flotation tank 15.

[0028] As shown in FIG. 1 , the pressurized flotation tank 15 stores the treated water W1 supplied from the coagulation tank 14. The pressurized flotation tank 15 generates air bubbles in the treated water W1, causing flocs 31 to float to the water surface. The flocs 31 that have risen to the water surface are removed from the pressurized flotation tank 15 to the sludge tank 16, then transferred to the dehydrator 17, where they are dehydrated. As a result, the flocs 31 become sludge cakes, which are discharged from the dehydrator 17. The flocs 31 that have become heavy due to coagulation settle to the bottom of the pressurized flotation tank 15. The flocs 31 that have accumulated at the bottom are then sent to the sludge tank 16 by a pump (not shown), then transferred to the dehydrator 17, where they are dehydrated. In this case, the flocs 31 become sludge cakes, which are also discharged from the dehydrator 17.

[0029] The pH adjustment tank 18 stores the treated water W1, from which flocs 31 have been removed and which has been discharged from the pressurized flotation tank 15. A flocculant F2 (caustic soda) is supplied from the second tank 22 to the treated water W1 stored in the pH adjustment tank 18. This adjusts the pH value of the treated water W1. The pH adjustment tank 18 then supplies the treated water W1, whose pH value has been adjusted, to the discharge tank 19, and the treated water W1 supplied to the discharge tank 19 is transferred to a wastewater treatment plant (not shown).

[0030] As shown in Fig. 1, an IP camera 40 (imaging device) is installed in the coagulation tank 14 to capture an image 41 (see Fig. 2(a)) by capturing an image of the treated water W1. The IP camera 40 then stores image data of the captured image 41. Note that the image 41 in this embodiment is a color image.

[0031] Next, the electrical configuration of the wastewater treatment management system 1 will be described.

[0032] As shown in FIG. 1, the wastewater treatment management system 1 includes a personal computer 50, which includes a control device 51 that controls the entire system. The control device 51 includes a CPU 52, a ROM 53, a RAM 54, an input / output circuit, and the like. The CPU 52 is electrically connected to the pressurized flotation tank 15, the dehydrator 17, the first tank 21, the second tank 22, the polymer flocculant dissolving tank 23, a pump (not shown), the IP camera 40, and the like, and controls them using various drive signals. A display 55 and a keyboard 56 are also electrically connected to the CPU 52. The display 55 in this embodiment is a touch panel display. An image 41 captured by the IP camera 40 is displayed on a display screen 55a (see FIG. 2(a)) of the display 55. The ROM 53 stores a program for controlling the wastewater treatment management system 1.

[0033] Next, a wastewater treatment management method using the wastewater treatment management system 1 will be described.

[0034] First, the CPU 52 outputs a drive signal to the IP camera 40 and controls the IP camera 40 to capture an image 41 of the treated water W1 in the coagulation tank 14. The IP camera 40 stores the captured image 41. Next, the CPU 52 performs processing of a detection model construction step and constructs a detection model for detecting flocs 31 from the image 41 of the treated water W1 stored in the IP camera 40. In other words, the CPU 52 functions as a "detection model construction means." The CPU 52 then stores the constructed detection model in the RAM 54.

[0035] If the constructed detection model is applied to the entire image 41, disturbance portions 42 (see FIG. 2(a)) of the image 41 may adversely affect the detection of flocs 31. The disturbance portions 42 are portions caused by, for example, the reflection of lighting or piping (not shown) passing through the coagulation tank 14. Therefore, in this embodiment, an analysis area 43 (see FIG. 2), which is an area for detecting flocs 31, is arbitrarily set by the administrator. The analysis area 43 is a rectangular area set in the image 41 excluding the disturbance portions 42. The analysis area 43 is set by the administrator by, for example, dragging and dropping the mouse on the display 55.

[0036] Furthermore, the IP camera 40 is basically operated as a surveillance camera. Therefore, the angle of view of the IP camera 40 when capturing an image of the treated water W1 is determined by the administrator changing the zoom (e.g., by touching a zoom change icon (not shown) displayed on the display 55). However, changing the zoom may significantly change the image quality of the image 41 captured by capturing the treated water W1 from the conditions used when the detection model was constructed. This may result in a decrease in the accuracy of detecting flocs 31, even if image processing is performed using the detection model. Therefore, in this embodiment, the CPU 52 automatically controls the zoom of the IP camera 40 when capturing an image of the treated water W1 so that the angle of view is the same as the conditions used when the detection model was constructed, and detects flocs 31 under this condition. This allows the IP camera 40 to function both as a surveillance camera and to capture images of the treated water W1. After detecting flocs 31, the CPU 52 automatically returns the angle of view of the IP camera 40 to the previous angle of view.

[0037] In the subsequent floc detection step, the CPU 52 performs image processing using a detection model in an arbitrary analysis region 43 previously set in the image 41, thereby detecting multiple flocs 31 within the analysis region 43. Note that flocs 31 outside the analysis region 43 are ignored. In other words, the CPU 52 functions as a "floc detection means." Note that each detected floc 31 is surrounded by a rectangular bounding box 44 (see FIGS. 2 to 4). Furthermore, the CPU 52 detects flocs 31 from the image 41 acquired every predetermined period (one hour in this embodiment), making it possible to confirm the generation process (transition) of flocs 31.

[0038] The CPU 52 manages data pairs each including the number of flocs 31 detected within the analysis region 43 (see "97 ct (count)" in FIG. 3(a) and "37 ct" in FIG. 4(a)) and the mode value of the length L1 of the long side 45 of the bounding box 44 surrounding the detected flocs 31. The length L1 of the long side 45 can be obtained by calculating the difference between the X coordinate of the first end of the long side 45 (the right end in FIG. 3(c) and FIG. 4(c)) and the X coordinate of the second end of the long side 45 (the left end in FIG. 3(c) and FIG. 4(c)). The data is data obtained for each generation process of the flocs 31 (see images "1," "2," "3," ..." in FIG. 5), and is recorded (stored) in the RAM 54 and managed as a history. In other words, the CPU 52 functions as a "data management means."

[0039] Next, the administrator (quantification means) quantifies the size of the flocs 31 based on the managed data. Specifically, the administrator performs a sensory evaluation by looking at the state of the image 41 stored as a history, and evaluates the size by substituting a numerical value for the size of the flocs 31 (see "Y" in FIG. 5). If the substituted numerical value is equal to or greater than the target length (20 mm in this embodiment), the administrator determines that the size of the flocs 31 is appropriate (see "Good" in FIG. 5). On the other hand, if the substituted numerical value is less than the target length, the administrator determines that the size of the flocs 31 is inappropriate (see "Not Good" in FIG. 5). Note that if the number of detected flocs 31 is less than a certain number, the individual flocs 31 in the analysis area 43 will be large, and the administrator can therefore estimate that the flocs 31 are of a size that is easy to remove from the treated water W1 (good quality). On the other hand, if the number of detected flocs 31 is equal to or greater than a certain number, the individual flocs 31 in the analysis area 43 become small, and the administrator can estimate that the flocs 31 are of a size that makes them difficult to remove from the treated water W1.

[0040] The CPU 52 then accumulates (stores in the RAM 54) the quantified size data of the flocs 31. The CPU 52 also records the results of the floc 31 size determination in a management form 61 and automatically distributes (displays on the display screen 55a of the display 55) the results (see FIG. 6). The CPU 52 also displays the status of the wastewater treatment facility 10, the progress of the determination results, the evolution of the images 41, and the like on the display screen 55a. Specifically, the display screen 55a displays a gallery 62 showing the status of the control panel (control device 51) that controls the wastewater treatment facility 10 at the top and a schedule for detecting flocs 31 at the bottom (see FIG. 7). The display screen 55a also displays a gallery 63 (see FIG. 8) showing the progress of the pH value and a gallery 64 (see FIG. 9) showing the progress of the images 41. The administrator can switch between the galleries 62 to 64 by operating the display 55 with a mouse, for example. The information displayed in the galleries 62 to 64 may be distributed using a dashboard, and the displayed content may be automatically switched.

[0041] In the subsequent prediction model construction step, the CPU 52 constructs a prediction model that predicts the size of the flocs 31 after a predetermined time (three hours in this embodiment) and the image 41 after the predetermined time by accumulating data on the size of the flocs 31 quantified by the administrator and monitoring data as a pair. That is, the CPU 52 functions as a "prediction model construction means." Note that the monitoring data in this embodiment is data including the amount of treated water W1 received by the coagulation tank 14, the amount of chemicals (coagulants F1 to F3) added, and the pH value of the treated water W1 in the coagulation tank 14. The CPU 52 then stores the constructed prediction model in the RAM 54.

[0042] The CPU 52 then performs optimization calculations using a prediction model based on the amount of received treated water W1, the current pH value, and the current size of the flocs 31. This calculates the dosage amounts of chemicals (flocculants F1-F3 and calcium ions) that will allow the size of the flocs 31 after a predetermined time (three hours in this embodiment) and the image 41 after the predetermined time to reach the target. That is, the CPU 52 functions as a "dosage calculation means." The CPU 52 then outputs a drive signal to the first tank 21, controlling the automatic supply of the calculated dosage amount of flocculant F1 from the first tank 21 to the first reaction tank 12. The CPU 52 also outputs a drive signal to the second tank 22, controlling the automatic supply of the calculated dosage amount of flocculant F2 from the second tank 22 to the second reaction tank 13. The CPU 52 also controls the automatic supply of the calculated dosage amount of polymer flocculant F3 from the polymer flocculant dissolving tank 23 to the coagulation tank 14. Furthermore, the CPU 52 performs control to automatically supply slaked lime (calcium ions) to the first reaction tank 12 in the calculated input amount.

[0043] The manager checks the report (management form 61) displayed (distributed) on the display screen 55a of the display 55 at predetermined intervals to check the size of the flocs 31 detected by the CPU 52 and the results of the automatic supply of chemicals by the CPU 52. The manager also visually determines the actual size of the flocs 31 contained in the treated water W1. The manager then determines whether the detected size of the flocs 31 deviates from the actual size of the flocs 31. If the manager determines that the detected size of the flocs 31 deviates from the actual size of the flocs 31, the manager corrects the data of the size of the flocs 31 detected by the CPU 52 to data of the size of the flocs 31 determined visually and inputs the corrected data. The data is input by the manager operating the mouse and the keyboard 56. The CPU 52 then automatically updates the prediction model based on the input data.

[0044] Therefore, according to this embodiment, the following effects can be obtained.

[0045] (1) In the wastewater treatment management system 1 of this embodiment, the detection model detects flocs 31 within an analysis region 43 set in an image 41 of the treated water W1, and the prediction model predicts the size of the flocs 31 after a predetermined time. This automates sensory evaluation (specifically, evaluation of the coarsening state of flocs 31), which was previously performed visually by the manager, and eliminates the need for the manager to look into the treated water W1, significantly reducing the burden on the manager. Furthermore, based on the size of the flocs 31 predicted by the prediction model, the amount of chemicals (flocculants F1 to F3 and calcium ions) added to coarse the flocs 31 can be minimized, which is economical.

[0046] (2) The IP camera 40 of this embodiment functions as a surveillance camera that monitors the coagulation tank 14, etc., during normal times when the treated water W1 is not being imaged. This eliminates the need to install the IP camera 40 that images the treated water W1 separately from the surveillance camera, thereby reducing the cost of the wastewater treatment management system 1.

[0047] The above embodiment may be modified as follows.

[0048] The CPU 52 in the above embodiment quantifies the size of the flocs 31 based on data that combines the number of flocs 31 detected in the analysis area 43 with the most frequent value of the length L1 of the long side 45 of the bounding box 44. However, the CPU 52 may quantify the size of the flocs 31 based on data that includes only the number of flocs 31 detected, or may quantify the size of the flocs 31 based on data that includes only the most frequent value of the length L1 of the long side 45.

[0049] In the above embodiment, the CPU 52 quantifies the size of the flock 31 based on data including the most frequent value of the length L1 of the long side 45 of the bounding box 44. However, the CPU 52 may quantify the size of the flock 31 based on data including the most frequent value of the length L2 of the short side 46 of the bounding box 44 (see FIGS. 3(c) and 4(c)), or may quantify the size of the flock 31 based on data including the most frequent value of the length of each side of the bounding box 44 (either the long side 45 or the short side 46).

[0050] In the above embodiment, a rectangular analysis region 43 is set in the image 41, but an analysis region having another shape, such as a triangle, square, diamond, trapezoid, circle, or ellipse, may be set in the image 41. Also, in the above embodiment, the flocks 31 in the analysis region 43 are surrounded by a rectangular bounding box 44. However, the flocks 31 may be surrounded by a bounding box having another shape, such as a triangle, square, diamond, trapezoid, circle, or ellipse.

[0051] In the above embodiment, the data stored in the prediction model were data on the size of flocs 31, data on the volume of treated water W1, data on the dosage of flocculant F1, data on the dosage of flocculant F2, data on the dosage of flocculant F3, and data on the pH value of treated water W1. However, the data stored in the prediction model need only include data on the size of flocs 31 from among data on the size of flocs 31, data on the volume of treated water W1, data on the dosage of flocculant F1, data on the dosage of flocculant F2, data on the dosage of flocculant F3, and data on the pH value of treated water W1.

[0052] In the above embodiment, the administrator digitizes the size of the detected flocs 31 using the CPU 52 based on the data managed by the administrator, i.e., the data obtained for each generation process of the flocs 31 (see images "1", "2", "3", ... in Figure 5). However, the digitization of the size of the flocs 31 may be performed by the CPU 52. In other words, the CPU 52 may have the function of "digitization means."

[0053] In the above embodiment, if the quantified size of the flock 31 is less than the target length (see "No" in FIG. 5), the CPU 52 may output a drive signal to the display 55 and control the display 55 to display an alert indicating that the size of the flock 31 is inappropriate (for example, the text "Flock is too small"). This notifies the manager that the size of the flock 31 is inappropriate.

[0054] In the above embodiment, the IP camera 40 integrated with a computer is used as the imaging device, but a camera without a computer may also be used as the imaging device. When a camera without a computer is used, the image 41 captured by the camera is stored in the RAM 54.

[0055] The IP camera 40 in the above embodiment has both a function as a surveillance camera and a function for capturing images of the treatment water W1. However, the IP camera 40 may be a camera installed separately from the surveillance camera.

[0056] In the above embodiment, the IP camera 40 that captures the treated water W1 and obtains the image 41 is installed in the coagulation tank 14 that stores the treated water W1. However, the IP camera 40 may be installed in the second reaction tank 13, which is another treatment tank that stores the treated water W1, in addition to the coagulation tank 14.

[0057] Next, in addition to the technical ideas set forth in the claims, the technical ideas grasped by the above-described embodiments will be listed below.

[0058] (1) A wastewater treatment management system according to any one of claims 1 to 6, wherein the analysis area is an area set in the image excluding disturbance areas.

[0059] (2) A wastewater treatment management system according to any one of claims 1 to 6, wherein the floc detection means detects the flocs from the image at predetermined intervals.

[0060] (3) The wastewater treatment management system according to claim 2 or 3, wherein the prediction model construction means constructs the prediction model for predicting the size of the flocs after a predetermined time and the image after a predetermined time by accumulating the data on the size of the flocs digitized by the digitization means and monitoring data as a set.

[0061] (4) In the technical idea (3), the monitoring data is data including the amount of the treated water received by the treatment tank, the amount of the coagulant added, and the pH value of the treated water in the treatment tank.

[0062] (5) The wastewater treatment management system according to claim 4, wherein the imaging device has both a function of imaging the treated water and a function of a monitoring camera.

[0063] (6) In claim 5, the wastewater treatment management system is characterized in that the input amount calculation means calculates the input amount of the coagulant that will allow the size of the flocs after a predetermined time and the image after a predetermined time to reach a target by performing optimization calculations using the prediction model based on the amount of the treated water received in the treatment tank, the current pH value of the treated water, and the current size of the flocs. [Explanation of symbols]

[0064] 1...Wastewater treatment management system 2...Painting equipment 10...Wastewater treatment equipment 14...Flocculation tank as a treatment tank 31...Flock 40...IP camera as an imaging device 41...Image of treated water 43…Analysis area 44...Bounding box 45...long side of bounding box 52...CPU as detection model construction means, floc detection means, prediction model construction means, digitization means, data management means, and input amount calculation means F1, F2...flocculants F3... Polymer flocculant as a flocculant L1: The length of the long side of the bounding box W1: Treated water W2…Drainage

Claims

1. A system for managing the state of treated water obtained by reacting a coagulant with the wastewater to generate flocs in a wastewater treatment facility that removes harmful substances contained in wastewater discharged from a coating facility by precipitating them as flocs, comprising: a detection model constructing means for constructing a detection model for detecting the flocs from the image of the treated water; a floc detection means for detecting the flocs in an arbitrary analysis region preset in the image by performing image processing using the detection model; a prediction model construction means for constructing a prediction model for predicting the size of the flocs after a predetermined time by accumulating data on the detected floc sizes; A wastewater treatment management system comprising:

2. a digitizing means for digitizing the size of the flocs detected by the floc detecting means, The prediction model building means accumulates data on the floc sizes digitized by the digitizing means. The wastewater treatment management system according to claim 1 .

3. a data management means for managing data pairs each including the number of flocs detected in the analysis area by the floc detection means and the mode of the length of the long side of a bounding box surrounding the detected flocs, The digitizing means digitizes the size of the flocs based on the data managed by the data management means. The wastewater treatment management system according to claim 2 .

4. The wastewater treatment management system according to claim 1 , further comprising an imaging device that is installed in a treatment tank that stores the treated water and captures an image of the treated water to obtain the image.

5. 2. The wastewater treatment management system according to claim 1, further comprising: an input amount calculation means for calculating an input amount of the flocculant that will result in the size of the flocs after a predetermined time by performing an optimization calculation using the prediction model.

6. 2. The wastewater treatment management system according to claim 1, wherein, when the size of the flocs detected by the floc detection means deviates from the size of the actual flocs contained in the treated water, the prediction model is automatically updated by correcting data on the size of the flocs.

7. A method for managing the state of treated water obtained by reacting a coagulant with wastewater to generate flocs in a wastewater treatment facility that removes harmful substances contained in wastewater discharged from a coating facility by precipitating the flocs, comprising: a detection model construction step of constructing a detection model for detecting the flocs from the image of the treated water; a floc detection step of detecting the flocs in an arbitrary analysis region preset in the image by performing image processing using the detection model; a prediction model construction step of constructing a prediction model for predicting the size of the flocs after a predetermined time by accumulating data on the detected floc sizes; A wastewater treatment management method comprising:

Images