Focused on the bearing detection, up nearly 18% accuracy

Focused on the bearing detection, up nearly 18% accuracy - 2019 04 - The future for 24 Han Jingxian industrial data flow entrance. When we're talking about the change from technology, mention is more of a consumer market, Internet of things, such as voice assistant, smart speakers. For the less 'mass' industry? From the point of economic common sense, the inefficient production mode inevitably was eliminated by way of higher efficiency, the current in machine learning, cloud computing and big data as the core of the new technology will bring a revolution for industrial production. Annual output value $6, detect employees accounted for a third 'why bearing areas? '' because in the field of manufacturing, bearing industry of automation and accuracy requirements are high, with bearing detection as a starting point, we then follow into other areas. 'depth vision technology director of sales &marketing Han He day in an entrepreneurial state in an interview. Bearing is a core infrastructure components in industrial manufacturing, global annual production in hundreds of billions, China now there are about 1500 bearing production companies, accounts for over 20% of the global market. Bearing detection is very important during the bearing production in a ring, a direct impact on the quality of the bearing and the bearing sales. However, for the detection of bearing product appearance mainly rely on manual. It is reported, China's largest bearing ring production enterprises, the annual output value $6, a total of 600 employees, including 200 individuals responsible for the test. Inspectors work eight hours a day to 20000 or so to check bearing, Han He day said: 'I tried once, I saw the tears in 10 minutes. 'along with the continuous renewal of testing equipment, also will be a part of the enterprise on the market at present defect detection using machine vision, but due to technical limitations, poor stability, basic error detection rate at about 20%, that is to say, the machine will detect 20% of good product as bad, it is must to use artificial for recheck, whether to process or for the cost, it is a huge waste. Bearing test and the contradiction between the current detection method, the importance of industry provides a breakthrough for the reform of machine vision technology, also offered the opportunity to enter for entrepreneurs. Machines instead of human eyes eyes, up nearly 18% accuracy for the low efficiency in the process of how to solve the bearing detection, low accuracy, slow speed, the problem of high labor costs, the depth of the visual technology from the angles of machine vision in intelligent integration testing machine to install multiple intelligent industrial camera, through the image capture, identification and calculation, playing the basic processes such as to complete the entire test. In the hardware part, depth vision for intelligent detection more than all-in-one mount 'eyes'. The first eye is visible light imaging, the second eye is intelligence long wave infrared cameras, metal surface layer 1 mm depth of crack can be detected, but also to do distinguish between scratches and third eye is the level of medical endoscope, put it in the bearing test the entire device, it can detect the aperture within 3 mm diameter internal scar. The developed industrial camera will add depth vision smart chip and the algorithm of independent development, let the camera have edge of computing power, that is to say, at the instant of the grab image, 90% of the amount of calculation is done within the machines' eyes', thus improve the operation efficiency of the equipment. , in enhancing the accuracy of depth perception among equipment joined the artificial intelligence algorithm, through deep study, accumulated about the defects of hundreds of styles, let equipment is more sensitive to the defects. After the technology from laboratory study, depth of visual spend about half a year of time to debug and experiment on factory production lines, to deal with the real production process factory of all kinds of situation. Han He day said: 'this is the other startups don't have resource advantages of the market. 'after the actual test in the factory, the depth of visual product list from the point of view of precision bearing detection, with Japan's best current detection equipment can only be compared to detect the defect of bearing ring 200 microns, the depth of the visual product direct driving up 100 times. Up nearly 18% on detecting accuracy, pure material cost to the customer to save at least 10%, at the same time a machine can replace the 12 artificial, solve the redundant testing of ills. The depth of the visual has landed in August, the first device according to Han He days, the top eight bearing manufacturing enterprise, there are five already got in touch with the depth of vision, the three have signed the order. In 2019 there will be about 20 domestic equipment will sell, probably intention to order there will be dozens of Taiwan. Future to do industrial data flow entrance in addition to bearing area, the depth of the vision of science and technology plan to testing equipment into the car, such as fasteners, auto precision machining, metal surface, as well as medical capsules need detection field, involved in the metal surface inspection and non-metallic surfaces. Han He day also said: 'the future market opened, we will accumulate a large number of products by machine vision image data, become industrial data flow entrance. Defects such as scar detection field, because of data statistics, the metal surface detection industry did not form a unified standard, but we can through the accumulation of massive databases can be of different sizes, different depth, different shapes of defects were analyzed, and the analysis of the causes of it, and then back to the front manufacturing, even formulate industry standards. 'source: 'the startup state' 10, 2018