【Introduction to the track】
Leverage the power of developers to explore the needs of diverse scenarios, develop a new computing network ecology, and explore diversified computing power network application scenarios for global developers, and encourage participants to develop and form rich computing power network innovative applications based on the capabilities of mobile cloud products .
【Schedule】
Note: The preliminary competition will be judged by the online expert system, the semi-finals will be an online roadshow, and the track finals and national finals will be carried out in the form of offline/online roadshows depending on the situation. If the competition schedule is adjusted, please pay attention to the official website of the competition, communication groups, emails and other notifications.
【Competition arrangement】
The competition is divided into five stages: registration, preliminaries, semi-finals, track finals, and finals. 15 teams advance to the preliminary rounds, 10 teams advance to the semi-finals, and the top 3 teams in the track finals are recommended to be selected for the national finals. Road Excellence Award. The teams that have been promoted in each review process should follow the unified arrangement of the competition to participate in the next round of competition review. If the team fails to participate due to personal reasons, it will be deemed as giving up the promotion quota for the next round, and the promotion quota (including bonus) will be postponed according to the ranking of this stage.
【Introduction to the competition】
1. The title of the preliminary competition
Lithium battery tab welding quality machine vision inspection
2. Background of the preliminary competition questions
Lithium-ion batteries have the advantages of high energy density, long cycle life, and low self-discharge rate. Tab welding is the key and difficult point in the entire lithium battery production process. At present, the commonly used tab welding is the ultrasonic method. Through the high-frequency vibration of the welding head, instantaneous high heat is generated, so that the tab and the cell foil produce intermolecular thermal fusion, so that there is a welding strength relationship between the two; if welding Insufficient strength will lead to a large internal resistance of the lithium battery, which will lead to the scrapping of the battery cell; if it is in a completely welded state, the small flow area will cause a large amount of heat to be generated during charging and discharging, resulting in safety risks such as the lithium battery catching fire. The detection of pole piece defects by machine vision has the advantages of no contact, no damage, and 100% detection, which is specifically reflected in: high accuracy, can accurately identify coating scratches, pole piece damage, and asymmetrical welding of tabs, etc., accurately The detection rate is much higher than that of human eyes; and it excludes the influence of human factors such as the subjective will, emotion, and visual fatigue of the test personnel on the battery test results, and has high reliability, fast speed, high efficiency and low cost.
3. Preliminary competition questions and tasks
The purpose of this competition is to use artificial intelligence technology to develop a set of lithium battery tab welding quality inspection algorithms, and use computer vision to replace human eyes for various quality, safety, and integrity inspections.
1. Dataset introduction
The data set is the welding pictures of lithium battery tabs marked by classification, and the classification includes "jssx metal debris (1), OK qualified (2), lj edge debris (3), ong other NG (4), tfhd carbon powder black spots (5)"5 categories, the pictures include training set and test set, and the training set and test set both cover the above five types of classification data. The training set data is open to the contestants, and the test set is used as reserved data for the test and evaluation of the model.
Note: For details of image examples of qualified samples and defective samples, please refer to the data download.
4. Contents of the semi-finals, track finals and national finals
There is no specific proposition, and the participating teams are required to start from the perspective of machine vision inspection of lithium battery tab welding quality, consider relevant application scenarios in the industrial field, and combine the existing product capabilities, mobile cloud capabilities, and mobile computing power networks of the OnePower industrial Internet platform. Create related systems and application solutions for the integration of cloud and edge architectures. It is hoped that participating teams can pay attention to the pain points of industrial enterprises and propose creative solutions.
【prize settings】
1. Awards for the track finals:
The top 3 teams in the track finals will be selected for the finals, and the 7 teams after the track finals will be awarded the Track Excellence Award.
Track Excellence Award: 7 teams
2. National final awards:
There are four sub-tracks in the industry track: industry application innovation sub-track (6 teams), medical industry application sub-track (3 teams), industrial industry application sub-track (3 teams), and cloud computer application sub-track (3 teams). 15 teams on the track will compete in the national finals, and the awards are set as follows:
First Prize: 1 team
Second Prize: 4 teams
Third Prize: 10 teams
The national finals are planned to be carried out offline. For the travel, board and lodging expenses of the participating teams, the specific number of reimbursements and reimbursement standards, please consult the assistant of the competition.
【Application Requirements】
1. Registration method
Use your mobile cloud account to log in to the official website of the competition, click the "I want to participate" button on the competition page to register, and complete real-name authentication when submitting your work (click "upper right corner-account center-real-name authentication). (If you have any questions, please check: Download- Mobile Cloud Cup Competition Registration Process Manual)
2. Form a team
Each person can only participate in one team in each track, and entries are submitted in teams. Players are invited to form teams freely, and each team is limited to 2 to 5 people. After successful registration, players can create a team in "My Team" or join an existing team. The player who creates the team is the captain. The captain cannot create two or more teams across the track at the same time.
3. Resource acquisition method
Use the Mobile Cloud Internet account to obtain resources through the Internet free trial policy.
【Work submission】
1. After the registration is completed, the contestants must submit their works before August 15
2. Specific requirements for submitting work materials
Requirements for the Preliminary Round - Algorithm Competition:
Contestants must provide:
1) Algorithm model;
2) detective script;
A function introduction. The predictor module is an inference script, which aims to load the trained model through this script, and obtain the analysis results of the samples to be analyzed in the test set according to the incoming test set data samples. The organizer will call the reasoning script program to obtain the specific experimental results and running time, and calculate the results through the evaluation script to obtain the evaluation score.
B input and output. The predictor is set as a class, and the contestant needs to specify the model saving location and load the model during initialization. When using the function predict in the class to perform inference operations, you need to set the incoming parameters (single image data path); the return value of the function should be the category number corresponding to the image in the test data set (jssx metal debris (1), OK qualified (2) , lj edge debris (3), ong other NG (4), tfhd toner black spots (5)).
The pseudocode of Predictor is described as follows:
import...
class Model(): # 主类
def __init__(self,p_path):
self.__model = None
# 使用按照压缩包路径下的相对路径，
# 如，解压后的文件夹名称为model,模型存放于model下的pysrc/model.pkl路径，则路径为'model/pysrc/model.pkl'
try:
# -------------------------如下位置可修改(加载模型)------------------------
self.load_model(p_path+'/model/pysrc/model.pkl')
# ------------------------- END ------------------------
except Exception as e:
...
return None
def predict(self,path_to_img:str): # 入参：单个图片的路径，如’aaa.jpg’
# Read in image
try:
# -------------------------如下位置可修改(读取图片)------------------------
x = Image.open(path_to_img)
# ------------------------- END ------------------------
except Exception as e:
...
# predict & score
try:
# -------------------------如下位置可修改(推理过程)------------------------
prediction = self.__model(x)
# ------------------------- END ------------------------
except Exception as e:
...
return prediction # 返回结果：整数，与label对应关系为：（jssx金属碎屑(1)、OK合格(2)、lj边缘碎屑（3）、ong其他NG（4）、tfhd碳粉黑点(5)）
def load_model(self,path_to_model:str):
# Load model
# -------------------------如下位置可修改（加载模型）------------------------
with open(path_to_model,'rb') as fp:
paras = pickle.load(fp)
self.__model = lambda x : random.randint(paras[0],paras[1])
# ------------------------- END ------------------------
return 0
调用Predictor类的伪码如下：
if __name__ == '__main__':
M = Model('./') # 初始化模型
print(M.predict('2023_zyy_35.jpg')) # 调用推理过程
# 输出：1-5之间的整数Note: The directory structure of the uploaded works folder is as follows (the sample code is for download, see Downloads):
├── model user upload directory
│ ├── __init__.py package functions
│ ├── predictor.py prediction function
│ └── pysrc model path
│ ├── model.pkl model (format can be changed according to the model)
The upload and related paths are subject to comments.
Semi-finals, track finals and national finals-creative competition
1) If the submitting party uses cloud resources for development and deployment, it must be based on the mobile cloud and run on the mobile cloud. The submission proposal includes but is not limited to: PPT for the introduction of the work with recorded audio explanations (required), work Demo/available Running code (required), work demonstration video (optional), etc., all content must be packaged in zip format as a whole, and the file size should not exceed 500MB. Zip package naming rules: rematch + team name + work name + team leader contact number.zip.
What problems and pain points are solved by A, scenario and function description (required)
Technical architecture and business model of project B (optional)
C Which products or services of the mobile cloud are used (required)
D Business prospect and application value (optional)
【Review Dimensions】
Preliminary Round - Algorithm Competition:
1. Judging criteria for competition questions
In this competition, contestants design and train models by classifying and marking the image data sets in advance. It is required that the trained model can be used in the test set to complete model reasoning. The final model can realize the identification of qualified pictures and defective pictures, and classify the defective pictures according to "metal debris, carbon powder black spots, edge debris, and other NG".
2. Scoring mainly considers two dimensions:
1) f1 value of image classification (f1);
The formula for calculating F1i of each category is as follows:
F1i = 2PR / (P+R)
-R (recall rate) = number of detected objects / (number of detected objects + number of missed objects)
-P (Precision Rate) = Number of Correct Targets Detected / (Number of Correct Targets Detected + Number of False Detections)
B is finally calculated using the F1i of each category (macro-F1):
f1= sum(F1i)/number of categories
2) The total inference time (t), the system limits the time to complete inference to be less than 180s.
3) The total score is weighted by the f1 value and the total reasoning time. The weight of the f1 value is 0.9, and the proportion of reasoning time is 0.1. The scoring formula is:
Semi-finals, track finals and national finals-creative competition
The selection of the creative competition is carried out from the following dimensions:
1. Innovation value
Whether there is any innovation in the thinking and method of solving the problem.
2. Commercial value
It can create market benefits and has broad market prospects.
3. Industry prospect
Whether it conforms to the status quo of the industry and solves the needs and pain points of the industry.
4. Functional integrity
The core functions are available and objectively feasible.
5. The logical process of product interaction is smooth and the user experience is good.
6. Trial situation of mobile cloud product capabilities
The rationality and tightness of the use of mobile cloud product capabilities.
7. Additional points
1) The work is closely related to the competition topic, starting from the general direction of "computing power network" and "calculating from east to west", and has a good idea.
2) Have the design or implementation concept of basic computing network strategies such as "computing network scheduling strategy", "traffic drainage strategy", "computing resource synchronization strategy" and "network resource synchronization strategy", and have the unified arrangement and coordination of "computing network brain" Scheduling design or implementation concept.
【Organizational Unit】
Sponsor: China Mobile Communications Group Co., Ltd.
Organizer: China Mobile Cloud Competence Center
Co-organizers: China Mobile (Shanghai) Information and Communication Technology Co., Ltd., China Mobile Communications Group Liaoning Co., Ltd., Inspur Communication Information System Co., Ltd.
【Handling of Violations】
The entries must be authentic, healthy, and legal. If the following violations are found, the competition organizing committee has the right to disqualify the contestants or participating teams:
1. Cheating or falsifying the registration information;
2. Behaviors suspected of plagiarism or infringement of intellectual property rights of others, directly applying open source codes, etc.;
3. Violation of relevant laws, regulations, policy provisions and relevant rules of the official website of Mobile Cloud during the competition;
4. The submitted works involve unhealthy or defamatory third-party content;
5. Use third-party AI software (such as ChatGPT, etc.);
6. Other violations found or reported to exist during the competition.
【Appeal and Arbitration】
1. The participating teams or contestants can appeal to the competition organizing committee for equipment, tools and software that do not meet the competition regulations, unfair judgments and rewards, and staff violations. The competition organizing committee is responsible for accepting the appeals raised during the competition and conducting mediation and arbitration, so as to ensure the smooth progress of the competition and the fairness and justice of the competition results. The arbitration result made by the competition organizing committee is final.
2. The appeal report should clarify the content of the appeal and provide relevant evidence. It is required that the appeal must be real-named, otherwise the appeal will not be accepted.
3. The competition organizing committee will accept the appeal report within 5 working days from the date of receipt, and carefully review and handle it.
【other】
1. During the course of the competition, there may be a small amount of changes and adjustments to the competition rules, and all content is subject to the official website of the competition.
2. The entire event will be photographed and videotaped; entries will be used for display and publicity on the official website of the mobile cloud developer community.
3. The final interpretation right of this competition regulation belongs to the organizing committee of the competition.
Competition details consultation :
