Introduction of reasonable fertilization method for fish ponds

You may use fertilizers directly and let them dissolve slowly. Water currents help to disperse the dissolved chemicals through the entire pond area. In small ponds one may use at least one fertilizing point per 1000 m2 of water area. In case of larger ponds , use 2-3 points per hectare, tentatively.

 

1. Retain the fertile water

When the pond is cleared in winter, the pond where the fish are grown should try to retain the fertile water of the original pond, so that once the fish species are put into the fish pond, there will be richer feed for consumption.

 

2. Avoid using raw fertilizers

The fertilizers used in fish ponds must be fermented, and raw fertilizers must not be applied.

 

3. Fertilize according to the weather and water quality

The amount of fertilization depends on the weather and water quality. In spring and autumn, the weather is cool, the water temperature is low, the decomposition of organic matter is slow, the oxygen consumption is low, the fish's food intake is small, the activity ability is poor, and the dissolved oxygen in the water is also large. Vigorous, the amount of fertilization should be small and many times. The specific method should be considered according to the conditions of the fish pond itself and the activity of the fish observed in the early morning in the pond, as well as the weather.

When the water quality is fertile, the weather is sultry, and the air pressure is low, most of the fish float their heads. At this time, the application should be stopped or less applied; if the water quality is thin, the weather is cool, and the fish are active in the water, more fertilizer should be applied.

If human and animal manure is used as fertilizer, it must be fermented to prevent the rapid fermentation of immature fertilizer at high temperature and a large amount of oxygen consumption, causing pond fish to float and die due to lack of oxygen.

 

4. Sprinkling the whole pond

Fertilizer should avoid the food court, adopt the method of sprinkling the whole pond, and avoid pouring the whole bucket of fertilizer into the pond.

 

5. Sewage should be tested first

If domestic sewage is used, it must be tested first, and then used after it is determined to be non-toxic.

 

6. Fertilization of fingerling ponds

The fertilization of fingerling ponds is slightly different from that of adult fish. After clearing the pond, irrigate the pond with water, and then apply a certain amount of fertilizer, in order to cultivate the food that the small fry can eat. 

The timing of fertilization must be mastered, either sooner or later. If it is applied early, the individual plankton will grow too large and the fry cannot be swallowed; if applied late, the fry will not be able to eat, which will affect the growth.

 

 

I.  The basic elements of the HMM model (Hidden Markov Model)

1. Five basic elements of HMM

In HMM, there are 5 basic elements: {I, O, A, B, π}. Let us have an introduction to these 5 basic elements in combination with the sequence marking task:

 

(1) I: state sequence. Here, it refers to the label behind each word.

(2) O: Observation sequence. Here, it refers to each word itself.

(3) A: State transition probability matrix. Here, it refers to the probability that a certain annotation will transfer to the next annotation.

(4) B: Observation probability matrix, that is, emission probability matrix. Here, it refers to the probability of generating a certain word under a certain label.

(5) π: Initial probability matrix. Here, refers to the initialization probability of each annotation.

in:

I = ( i 1 , i 2 , i 3 . . . i N ) I = ({i_1,i_2,i_3...i_N})I=(i

1

​

 ,i

2

​

 ,i

3

​

 ...i

N

​

 ) state sequence

O = ( o 1 , o 2 , o 3 . . . o N ) O = ({o_1,o_2,o_3...o_N})O=(o

1

​

 ,o

2

​

 ,o

3

​

 ...o

N

​

 ) observation sequence

A = [ a i j ] N ∗ N A = [a_{ij}]_{N*N}A=[a

ij

​

 ]

N∗N

​

  State transition probability matrix

B = [ b j ( k ) ] N ∗ N B = [b_j(k)]_{N*N}B=[b

j

​

 (k)]

N∗N

​

  Observation probability matrix

π initial state probability vector

 

2. HMM model

 

Model: λ = (A,B,π)

A, B, π are the three elements of the Hidden Markov Model

These three elements are obtained through statistics, these three values ​​are the parameters of the model, and the process of obtaining these three values ​​is the process of model training, so the HMM algorithm is a relatively simple algorithm. 

The model has been known, and it can be considered that the fully connected path of each state has been known. Given an observation sequence, the optimal path among all paths is solved by the veterbi algorithm.

 

3. Two assumptions of the HMM model

(1) Homogeneous Markov assumption (also called first-order Markov assumption)

The state of the hidden Markov chain at any time t only depends on the state at the previous time, and has nothing to do with the state at other times and the observed state.

P ( i t ∣ i t − 1 , o t − 1 , . . . , i 1 ) = P ( i t ∣ i t − 1 ) P(i_t|i_{t-1},o_{t-1},..., i_1) = P(i_t|i_{t-1})P(i

t

​

 ∣i

t−1

​

 ,o

t−1

​

 ,...,i

1

​

 )=P(i

t

​

 ∣i

t−1

​

 )

(2) Observational Independence Hypothesis

The observed state at any time only depends on the state of the Markov chain at that time, and has nothing to do with the observed states at other times.

The above elements can be counted from the training corpus. Finally, based on these statistics, we apply the Viterbi algorithm to calculate the label sequence behind the word sequence.

 

2. There are three application scenarios for the Hidden Markov Model

We only use one of them for named entity recognition - to find the most likely label sequence behind the observation sequence.

Finally, let's talk about the three problems that HMM solves:

 

1. Evaluation (probability calculation problem)

Knowing the model parameters λ= (A, B, π), calculate the probability of a certain observation sequence, that is, find P(O|λ)

 

2. Learning (learning problems)

Given a sequence of observations O = ( o 1 , o 2 , . . . , o n ) O=(o_1,o_2,...,o_n)O=(o

1

​

 ,o

2

​

 ,...,o

n

​

 ), how to adjust the model parameters λ=(π, A, B) to maximize P(O|λ)? , this is the algorithm to find the model

 

3. Decoding (prediction problem or decoding problem) is the most commonly used

Given an observation sequence O and a model λ, find the most probable state sequence S(s1,s2,…st+1).

For example: through entity labeling and training, we can get the model λ, now given an observation sequence = I work in Phoenix Finance, to find the most likely named entities, and want to find the corresponding state sequence S = (I, in , Phoenix Finance, Work).

 

3. Find the model λ: solve the second problem

HMM is a generative model, so the joint probability is sought

 

Note: When we usually say, finding the model refers to finding the objective function. For example, in linear regression, our objective function is $h(λ)=λ_1X+λ_2$, and finding the objective function only requires the parameter λ, so usually We say that seeking model is seeking parameters.

1

P ( X , Y ) P(X,Y)P(X,Y)

 

Fourth, Viterbi algorithm (Viterbi): solve the third problem

The Viterbi algorithm mainly uses dynamic programming to solve the prediction problem of HMM: always model and observation sequence, find the most probable state sequence

 

Suppose the sequence of states is: x 1 , x 2 , x 3 . . . x N x_1,x_2,x_3 ... x_Nx

1

​

 ,x

2

​

 ,x

3

​

 ...x

N

​

 

The corresponding observation sequence is: y 1 , y 3 , y 3 . . . y N y_1,y_3,y_3 ... y_Ny

1

​

 ,y

3

​

 ,y

3

​

 ...y

N

​

 

Then our problem is transformed into: the known input sequence y 1 , y 3 , y 3 . . . y N y_1,y_3,y_3 ... y_Ny

1

​

 ,y

3

​

 ,y

3

​

 ...y

N

​

 , corresponding to the most probable Chinese characters x 1 , x 2 , x 3 . . . x N x_1,x_2,x_3 ... x_Nx

1

​

 ,x

2

​

 ,x

3

​

 ...x

N

​

 . What is the most likely sequence of Chinese characters?

formula:

x 1 , x 2 , x 3 . . . x N = A r g M a x P ( x 1 , x 2 , x 3 . . . x N ∣ y 1 , y 3 , y 3 . . . y N ) = A r g M a x ∏ i = 1 N P ( y i ∣ x i ) ∗ P ( x i ∣ x i − 1 ) x_1,x_2,x_3 ... x_N = ArgMaxP(x_1,x_2,x_3 ... x_N|y_1,y_3,y_3 . .. y_N) = ArgMax \prod_{i=1}^N P(y_i|x_i)*P(x_i|x_{i-1})x

1

​

 ,x

2

​

 ,x

3

​

 ...x

N

​

 =ArgMaxP(x

1

​

 ,x

2

​

 ,x

3

​

 ...x

N

​

 ∣y

1

​

 ,y

3

​

 ,y

3

​

 ...y

N

​

 )=ArgMax∏

i=1

N

​

 P(y

i

​

 ∣x

i

​

 )∗P(x

i

​

 ∣x

i−1

​

 )

 

where the formula A r g M a x ∏ i = 1 N P ( y i ∣ x i ) ∗ P ( x i ∣ x i − 1 ) ArgMax \prod_{i=1}^N P(y_i|x_i)*P(x_i|x_{i-1 })ArgMax∏

i=1

N

​

 P(y

i

​

 ∣x

i

​

 )∗P(x

i

​

 ∣x

i−1

​

 ) is mainly transformed by the Bayesian formula

 

We know that the Bayesian formula is: P ( A ∣ B ) = P ( B ∣ A ) ∗ P ( A ) P ( B ) P(A|B) = \frac {P(B|A)*P(A )}{P(B)}P(A∣B)=

P(B)

P(B∣A)∗P(A)

​

 

Then P ( x ∣ y ) = P ( y ∣ x ) ∗ P ( x ) P ( y ) P(x|y) = \frac {P(y|x)*P(x)}{P(y) }P(x∣y)=

P(y)

P(y∣x)∗P(x)

​

 , where P(y) is a known constant, where P(x) is actually P ( x t ∣ x t − 1 ) P(x_t|x_{t-1})P(x

t

​

 ∣x

t−1

​

 ), according to the Markov hypothesis, the current moment hypothesis is related to the previous moment.

 

For example, enter the observation sequence:

I love China

O O O B

O B O I

O O B I

B O I B

That is, the third line sought is the optimal path:

 

Fourth, the Viterbi algorithm (Viterbi)

Note: During the calculation of the viberbi algorithm, the shortest path between two points is calculated, not the shortest path between the two layers.

1

1. Nature

If the path p with the highest probability (or the shortest path) passes through a certain point, such as a on the way, then the sub-path Q from the starting point S to a on this path must be the shortest path between S and X22.

Otherwise, replacing Q with the shortest path R from S to a constitutes a shorter path than P, which is obviously contradictory. It is proved that the optimality principle is satisfied.

 

2. Algorithms

If you find the shortest path between S and E, what better way than to traverse all the paths?

 

In fact there must be a shortest path among all paths:

 

 

Let's start solving step by step from scratch:

(1) First, the starting point is S, and there are three possible paths from S to A column: S-A1, S-A2, S-A3, as shown in the following figure:

 

 

We cannot arbitrarily say which segment of S-A1, S-A2, and S-A3 must be part of the global shortest path. So far, any segment may be an alternative to the global shortest path.

(2). Then start the second layer

<1> We start with the first node of the second layer:

 

As shown above, there are only 3 paths through B1:

 

S-A1-B1

 

S-A2-B1

 

S-A3-B1

If the final second layer node passes through B1, the shortest path must be selected from these three paths, then the other two can be deleted.

<2> Then we start the second node of the second layer:

 

 

Similarly, as shown in the figure above, there are 3 paths through B2:

 

S-A1-B2

 

S-A2-B2

 

S-A3-B2

If the final second layer node passes through B2, the shortest path must be selected from these three paths, and the other two can be deleted.

 

<3> Then we start the third node of the second layer:

 

Similarly, as shown in the figure above, there are also 3 paths through B3:

 

S-A1-B3

 

S-A2-B3

 

S-A3-B3

If the final second layer node passes through B3, one of the shortest paths must be selected from these three paths, and the other two can be deleted.

<4> After all the stages of the second layer are traversed, there are three paths left.

 

We don't yet have enough information to tell which one must be a subpath of the global shortest path.

(3) Then we continue the algorithm at the third layer:

 

We don't yet have enough information to tell which one must be a subpath of the global shortest path.

(4) Then we continue the algorithm at the last layer:

 

Point E is already the end point. We can know which one is the shortest path by comparing the total length of the three paths.

 

 

Abstract: With the rapid development of agricultural information intelligence, the traditional mode of relying on the sky for food has evolved into a new agricultural mode in which intelligent equipment monitors the production process and performs automatic regulation. How to process the data collected by these smart devices, especially the image data, extract useful information from it, and combine it with the actual production control measures, so as to promote the efficient and rapid operation of agricultural production, of great significance. 

This study applies image processing and machine learning to automatic disease identification and leaf age detection of rice, which is more efficient and accurate than traditional manual identification and diagnosis methods. 

The main contents include: First, the basic theories of machine learning algorithms and image processing are introduced, and corresponding theoretical guidance and technical solutions are provided for the subsequent research on disease identification algorithms and leaf age diagnosis algorithms. 

Then, in the research on the algorithm of rice disease identification, the machine learning method is used for rice disease identification. This paper mainly studies three common diseases of rice, namely rice blast, bacterial blight and bacterial leaf spot. 

The specific contents include: First, preprocess the rice disease images, then segment the rice disease spots, and establish the corresponding rice disease image set. 

Then, according to the pathological external manifestations of different lesions, multiple features were extracted, and the extracted features were optimized by principal component analysis. 

Afterwards, BP neural network and support vector machine were used to establish models, and the optimized features were classified and tested, and the BP neural network with higher recognition accuracy was selected as the final classification model. 

Finally, the improvement of the BP classification model is proposed, and the genetic algorithm is used to optimize the initial selection process of the weights and thresholds in the BP algorithm. The experimental results show that the improved GA-BP algorithm is feasible and improves the accuracy of disease identification. 

Then, in the research of the algorithm for diagnosing the leaf age of rice, the traditional image processing method was used to extract the veins of the rice leaves. 

The judgment method is used to extract the rice leaf veins, and the improved leaf vein extraction algorithm can successfully extract the main leaf veins of rice. And the algorithm was used to verify the diseased rice leaves and the disease-free rice leaves respectively, and the algorithm extraction effect was good. At the same time, based on the algorithm, combined with the leaf vein bias method, the leaf age diagnosis was realized through practical tests. 

Finally, combined with the results of rice disease identification and rice leaf age diagnosis, on the basis of identifying diseases, taking leaf age diagnosis technology as the core, aiming at different diseases, and then taking corresponding management measures to paddy fields according to different leaf age stages in time to carry out Smart regulation of rice in cold regions.

 

Chinese Abstract: With the improvement of computing power of computers and chips, research in the field of artificial intelligence has re-emerged. Among them, Convolutional Neural Network (CNN) is widely used in image recognition in various fields, and is gradually becoming more and more mobile and embedded.

Rice disease detection has always been the primary problem in plant disease detection. How to accurately and efficiently detect rice diseases and improve rice yield is a hot issue in agriculture. 

Based on artificial intelligence theory, the rice disease detection system is studied. The system is mainly designed from three parts: image acquisition, temperature and humidity acquisition, and disease detection model. 

The ARM is used as the main control unit of the embedded platform to realize the image acquisition of rice disease spots and the detection of environmental temperature and humidity. The collected images are used as the input of the detection model to complete the offline detection of rice diseases, and the diagnosis is assisted by environmental factors such as temperature and relative humidity. 

Model design on the PC, build the TensorFlow deep learning framework, complete the design and training of the rice disease detection model through the convolutional neural network, and transplant the optimized model to the embedded terminal to realize the detection and recognition function of the terminal. 

The test shows that the rice disease detection system can complete the basic functions. The image acquisition rate is 25 frames per second, and the resolution of the acquired disease spot image is 640×480. The humidity range measured by the system is 0-99.9% RH, and the temperature range is -40-80℃. Through test verification, the average accuracy of disease identification is 96%.

How to raise economic benefits for professional chicken raising households?

 

1. Selection of fine breeds

In chicken production, improved breeds are the basis for improving commodity rates, reducing production costs and improving economic benefits. 

Professional households should determine whether to raise broiler or laying hens according to the local market supply and demand and their own conditions, and which breed to choose.

 

2. Choose a good chicken

The quality of chicks is the basic factor to obtain better chicken raising benefits. Chicken farmers must choose young chickens provided by breeder farms with good reputation, certain production scale and high technical level, and drink water with 0.01% concentration of enrofloxacin hydrochloride at the first week of age to eliminate Salmonella, mold and mildew. 

Plasma and other pathogenic microorganisms, and do a good job in disinfection, isolation and epidemic prevention at the same time.

 

3. Self-prepared feed

Professional households can formulate full-price compound feeds that meet the needs of different physiological stages of chickens according to the feeding standards, or formulate their own formulas according to the formula provided by technicians, which generally reduces the cost of feed per 100 kg by 10 to 20 yuan.

 

4. Add yeast

Appropriately add high-quality yeast feed to reduce the amount of imported fish meal added to the feed. 

Generally, the nutritional value of each kilogram of active yeast powder is not less than 1 kilogram of imported fish meal, and it is about 1.6 yuan cheaper per kilogram. 

For example, the "Anying" brand "ESB probiotic yeast powder" produced by Xuzhou Anying Biotechnology Development Co., Ltd. can greatly enhance the digestive ability of the stomach and the liver's disease resistance and immunity after the chicken eats it, and quickly improve the conversion rate of protein feed. 

Increase the egg production rate and effectively reconcile the survival of cells in the chicken, thereby significantly improving the economic benefits of raising chickens.

 

5. Improve feeding equipment and reduce feed consumption

i. Improve the design of the trough According to the age of the chickens, choose the appropriate size trough with eaves, and the height of the trough should be 2 cm higher than the back of the chicken. 

ii. Change from flat to cage The flat rearing method occupies a lot of houses, and the feed is wasted seriously. Cage chickens live in cages, with less activity and less material consumption. 

iii. The temperature of the chicken house should be moderate It is necessary to keep warm and cold in winter, and to prevent heatstroke in summer, especially in winter, the temperature of the chicken house should not be lower than 10 ℃. 

If the ambient temperature is too low, the chicken body should produce more heat to maintain body temperature, so the consumption of feed is high and the feed payment is low.

 

6. Disease control

1. Antibody monitoring Conditional chicken farmers should carry out antibody monitoring, so as to ensure timely vaccination and the quality of immunity.

 

2. Strictly grasp the quality of vaccines Purchase vaccines from vaccine production and supply units with reliable quality. Strictly prevent vaccine production and supply units from purchasing vaccines. Strictly prevent the confusion of vaccines, otherwise the antibodies will be unstable, and repeated immunizations will be required.

 

3. Strictly follow the immunization program A strict and scientific immunization plan should be formulated according to the actual situation, and the prevention and control of chicken Marek's disease, chicken Newcastle disease, chicken Gan Paul's disease, chicken cholera and chicken pox should be done well.

 

4. Add antibiotics to the diet to prevent infection Such as antibiotics or sulfonamides, the prevention of diarrheal diseases (such as pullorum) is effective, but attention should be paid to the dosage and stirring, and the feeding time should not be too long.

 

5. Strengthen the preservation of feed to prevent diseases caused by mildew of feed.

 

7. Feeding and management

1. Regularly disinfect the chicken coop and environment to keep the chicken coop dry and hygienic. 

Regular deworming, free-range chickens are generally dewormed once a week, and broad-spectrum, high-efficiency, and low-toxic deworming drugs are used.

 

2. Eliminate the rooster in time. During the brooding stage, males

 

The chick consumes about 25% more feed than the mother chick. If it is not eliminated in time, it will not only consume a lot of material, but also occupy a lot of feeding area, and it will also affect the normal growth and development of the hen, so it is best to eliminate it when the chicks are just hatched. 

The male-female ratio of breeder flocks should be appropriate. The male-female ratio of egg breeders and broiler breeders should be 1:15-20 and 1:10-14 respectively. 

If artificial insemination technology is used, the ratio of male to female can be increased to 1:30 to 50, thereby reducing material consumption.

 

3. Pay attention to the observation of the flocks, and promptly eliminate sick or low-yielding chickens. In general, these chickens account for 5% to 8% of the total, and as many as 11%, they should be discovered and eliminated in time.

 

4. Provide proper lighting. The suitable light time for laying hens is 16 hours. For free-range laying hens, add half an hour to 16 hours of light per week from 18 weeks of age. The power of the bulb is 2.7 to 3 watts per square meter (in practical applications, the lamp height is usually 2 meters, and the lamp distance is about 3 meters). 

Pay attention to turn on and off the lights on time, and keep the lighting hours and brightness constant.

 

5. Timely beak cutting can not only prevent flock beak addiction, but also effectively reduce feed waste. Beak cutting and beak trimming time were 6-9 days and 12-17 weeks, respectively.

 

6. The "all-in-all-out" feeding management system is adopted to facilitate disinfection, disease prevention and scientific management.

 

8. Group feeding

Group feeding is one of the important measures for chicken production. As a feeding and management unit, a suitable group should be adapted to its own facilities, feed and technical measures. The flock of broiler chickens should not be too large. 

Generally, a group of about 400 chickens should be grouped, and the group should be divided into strong and weak groups from time to time. Stocking density of 10 to 15 per square meter is more appropriate. 

Breeding male and female hens in groups is not only conducive to the growth and development of hens, improving uniformity and commodity rate, but also better adapting to the market. 

The rooster group grows fast and has a large body weight. It can be slaughtered 7-10 days earlier, and the hens can be slaughtered later. Selling it as a whole chicken for cooking and eating can significantly improve economic benefits.

 

9. Selling at a high price

The market price of broilers at the time of slaughtering has a great impact on the economic benefits of raising broilers. Therefore, breeders should always understand market information, and determine the time and quantity of chickens according to the local chicken seedlings and the demand for feather chickens. 

When you are rushing up, you should raise less or not. When everyone stops raising, you should raise more, so that you can get great benefits. But be careful not to blindly wait for the price to rise to prolong the age of the flock, which is not cost-effective, because after 8 weeks of age, the feed return of broilers decreases, the mortality rate is high, and the overweight may not be easy to sell. Therefore, according to the market conditions and your own actual situation, you should be listed on the market in a timely manner.

 

10. Reduce expenses

Farmers need to keep their chickens simple and simple. They don’t need to build reinforced concrete houses. They can use plastic greenhouses with less investment. 

The use of water, electricity and drugs should be reasonable, and various appliances and equipment should be maintained to prolong their service life, reduce unnecessary waste, and do everything possible to reduce the cost of raising chickens, thereby improving economic benefits.

See also:

Feed of Commercial Broiler Chicken

Broiler Body Weight