Development of Conventional Controller Based on Image Processing for Monitoring and Controlling Burning Zone Temperature in a Cement Plant in Rotary Kiln Process Through IOT

Today, individuals' security demands are continually and mechanized robotization frameworks quality and quick reaction, considering his place in his life the plan of mechanical frameworks has gotten key. Article following, particularly at cement packing, Auto temperature finding in revolving kiln and item following focuses recognition in robot applications and curve welding robots at a steady speed of focuses and articles that should be development and change speed unpredictably. The cycle is getting more diligently. Open-close with sensor approach thus a more minimized route as opposed to framing an administrative body PID of the item wanted to be distinguished by picture handling It is more delicate to catch up with the review body and It produces dependable outcomes [1-3]. PID regulator examines were first begun in the naval force in 1922.Minorsky for programmed directing frameworks on ships It was applied by Minorsky [4]. PID controller studies were first started in the navy in 1922.With the simple structure of PID most of the industries adapting. Due to its simple structure of PID, it is widely used in different fields. Temperature detection with image processing, different studies have been done and these are applied in various fields to control speed, stability, error. It has formed its usage fields in the industry, in this method PID based image processing is used to measures the temperature using burning flame image in which it is considered Red, Green, Blue (RGB) magnitude planes. The system takes RGB magnitude as input variable and generates temperature as output. The proposed PID controller detects temperature of flame zones using color features of burning flame images, precise temperature mapping of flame images helps to control the temperature inside the rotational kiln to produce high quality clinker. This process can be viewed remotely and controlled using various control loops from anywhere.

OpenCV, created by the Intel firm Berkeley It is an open source programming authorized by Software Distribution (BSD). It is the library. Various PC vision calculations including. OpenCV library, C and C ++ by Intel created in programming dialects. Macintosh, Windows, Linux and so forth. It can chip away at stages [5]. Object investigation, military applications, security frameworks, clinical picture preparing applications, satellite Often OpenCV in picture and guide applications library is utilized. OpenCV form 2.3.1 is utilized in this examination. what's more, from its source code on the Raspberry Pi advancement board It has been ordered. Again, the use of OpenCV library in this study to Set the color values of the object to be detected as a threshold [5-11].

1. Pre-processing

The ideal examination of dependability is expanded by picture preprocessing systems. Typically every picture contains low recurrence commotion, so to seize undesirable data (clamor) picture needs to go through preprocessing stage [12]. The preprocessing is needed to improve the subtleties in picture. The picture upgrade changes the visual effect of the picture and improves the picture quality. In this work the upgrade is conveyed by middle channel.

2. Median Filtering

It is a nonlinear filter which preserves the edges of an image by confiscating the low frequency noise [13]. The random noise within communication channel generates a salt and pepper noise. It is a simple sliding spatial window that substitutes the middle value with average value of all pixels in the window. The window is usually a square kernel. The operation of filter is defined by the Eq. (1)

median [M(s)]+median [N(s)]      (1)

The signal processing procedure is able to perform the noise reduction in an image to get good perception. The objective of this filter is to sprint the image with pixel by pixel and replacing the each and every neighboring pixel with its median value. The configuration of neighbors is known as “window” which slithers pixel by pixel in entire image.

3. Segmentation

Region of picture into its constituent areas is known as division. It just changes the picture portrayal. Also, it is ordinarily used to limit articles and boondocks in pictures [14]. Ordinarily it is performed with various methods, for example, thresholding [6], Region based or locale developing and some other wavelet changes. Yet, all the division methods don't fragment the influenced part precisely. So in this work to identify or section the influenced part the cuckoo inquiry advancement method alongside thresholding is utilized as division apparatus.

4. Feature extraction

Highlight extraction is where it creates a lot of highlights otherwise called vectors. It depicts the substance of the picture and establishes the portrayal of the given picture. The capacity of Local Binary Pattern highlight extraction is to blackmail the highlights of relating area of worry in a picture.

5. Classification

This refers to carrying out data classes from a raster image with multiple bands. For creating topical maps the following raster from the image classification may be used. There are two types of classifications which depend on the cooperation between the expert and the PC in classification, namely supervised and Unsupervised. In this work CNN is used as classifier.

6. Feature extraction

Feature extraction is a process where it generates a set of features also known as vectors. It describes the contents of the image and constitutes the representation of the given image. The function of Local Binary Pattern feature extraction is to extort the features of corresponding region of concern in an image

From the caught digital frames of flame singular pixels can be isolated utilizing region developing system. Pixel area and its Red (R), Green (G) and Blue (B) colour [11] properties will be acquired for singular pixels. A color histogram can be utilized to speak to the colour arrangements of a picture.

Image processing calculation with Open CV shows up in the diagram. According to the algorithm, the frame value of the camera is setted. Conversion from RGB color space to HSV color space is achieved and the reference color threshold value is determined. Parasitic pictures are impeded by morphological sifting. Consuming zone pictures are recognized and their temperature is resolved (Figure 2).

2.png

Figure 2. Burning image co-ordinates

From Table 1 different images are giving various temperature values with error. By comparing the input image with set point we can judge which image is the best for giving proper clinker output. With the help of Table 1 we can control the raw meal feed rate, coal feed rate, % oxygen.

Table 1. Burning zone flame images with temperature and error value

Image	Measured Temperature (Deg –centigrade)	Error
650.png	650	750
750.png	750	650
1350.png	1350	50
920.png	920	480
1100.png	1100	300
1243.png	1243	157
1300.png	1300	100
850.png	850	550
1370.png	1370	30
1389.png	1389	11

In the catch of the article, the item sweep check is the info esteem. The pixel value which is esteem estimated by object following is contrasted and reference range. Range contrast is viewed as an error. Finding and controlling of temperature is performed by PID control as indicated by the determined error share and the directions of the article. These tasks proceed until the right image is caught to get the proper clinker ouput. In this we have taken 10 images randomly and measured the temperature and catch the right flame image to maintain constant temperature of 1400 deg centigrade approximately to get the proper clinker output.

A. PIDDC Motor Control

PID stands for proportional, integral and derivative controls. Because of its straight forward structure, the business is broadly utilized. PID acknowledges the distinction between the desired value and the reference an incentive inside the cycle as an error. It controls the cycle contribution to limit [8] the error value. The range esteem estimated by the camera is contrasted and the reference esteem. In this examination, the temperature of the burning zone is measured more proficiently by PID control. Implementation process of image processing technique detecting and controlling of temperature is shown below (Figure 3) [15, 16].

3.png

Figure 3. Temperature monitoring and controlling burning Diagram

B. Result image processing pi introduction to algorithm

Images which taken from the camera requested reference point is the pixel d in the middle of the camera value is image center of gravity calculated by matching [17, 18] and camera search for column value of middle pixel the difference as PI to be applied to the input of the controller an error value form. Feed back PI control systems in a digital environment when requested to apply PI [10] inspection discrete signal must be processed in time. This discrete in study time PI control signal set point 1423 deg centigrade. At this temperature proper clinker output will be produced.

4.png

Figure 4. The results of the Temperature example PID consisting of control signal

Figure 4 is an example case, followed move the subject to the right in front of the camera PI the controller will produce to fix the error the signal it emits is a section is given as an example. Here 1480 is the largest overshoot.

Example of split raise according to the man PI control signal, tracking in control mechanism of the object in front of the camera to keep the position stable 12v at 2000rp.Here taking 10 images at once and given the results. Results of the tracking temperature to the right of the midpoint in the camera time around taken from the slice.

Table 2. The tracked temperature of burning zone with percentage of error

From the data base for every 10 images for a second, during this time stamp different images are giving different temperature values. Actual temperature to get proper clinker was kept in set point, example set point is 1400 deg centigrade, this set point was get compared with the actual captured images.so that getting big error (Table 2).