The infrared image of solar cell's electroluminescence (EL) is one of the important means of hidden defects detection. In order to improve the automatic recognition rate of defect images, this paper adopts improved invariant moments for feature extraction. The scale factor of the improved invariant moments is eliminated by transformation. Therefore they have the properties of translation, rotation and scale invariance simultaneously in discrete state. At the same time, Support Vector Machine (SVM) is used to distinguish the defect image. The system which combined invariant moments with SVM is applied to classify the debris, crack, off-grid, open weld and black pieces. The recognition rate of 5 kinds of defects has reached more than 90%.

In order to improve the efficiency of electrostatic precipitators, a control system of constant current source based on Single Chip Microcomputer (SCM) is designed. In the control process, we use the pulse signal from microcontroller P1 port as digital signals of the switch. By changing the number of contact closure control, we achieve steady flow output of different magnitude. By using closed-loop feedback control, we achieve constant monitoring of dust concentration and ensure it compliance with outlet concentration of national requirements. The results show that microcontroller control system has a strong anti-interference, high accuracy, ease of programming, etc. Used in the electrostatic precipitator, collection efficiency can be further improved.

Automatic identification of defects on solar cell modules is the key of improving components efficiency of power generation and achieving the maximum utilization of energy[1]. This paper took the debris defects as an example. Considered the fact that scale factor would have an impact on moment invariants in discrete state, this paper adopted an improved moment invariant for the image feature extraction of debris defects. The scale factor of the new moment invariants is eliminated by transformation. Therefore they have the properties of translation, rotation and scaling simultaneously. Meanwhile, SVM (Support Vector Machine) classifier was used to identify the debris defects. The experimental results show that, compared with the traditional moment invariants, the identification rate of using the improved moment invariants for feature extraction is 96%.

In order to seek a better way of control in sewage treatment, a control system which consisted of water unit, sand unit, the oxidation ditch unit, precipitation unit and sludge dewatering unit was designed. The paper investigates various controllable parameters in sewage treatment system and the way of present common automation control. We designed manual control variable aeration program and two automatic control subroutine which included thickness grille and pumping station by using WIN 32 software. The program is designed in Siemens STEP 7 V5.5 software development platform and chose the 10 input contacts which are I0.0~I0.7, I1.0~I1.1 and 10 output contact which are Q1.0~Q1.1, Q0.0~Q0.7.This design achieves automatic control and improves the purifying efficiency on the premise of ensuring water quality.

In order to improve the automation level of the electrostatic precipitator, we used Very-High-Speed Integrated Circuit Hardware Description Language (VHDL) language to compile, emulate and optimize the control system of power source and vibration. In the design of Quartus platform, we used EP1C3T144C8 Field Programmable Gate Array (FPGA) chip to realize high voltage power supply, alarm and protection system. We also realized the compilation and simulation test of each part’s function of 20 s rapping and 40 min rapping cycle. At the same time, we recorded the waveform of simulation. It demonstrated that the validity of the relevant VHDL compilation. We used this method to achieve the optimization control of the electrostatic precipitator operating parameters. It has a strong practicability.