Single-Photon Avalanche Diode (SPAD) array detectors are commonly used for the ranging and three-dimensional imaging of long-range targets in space. In the application of detection with SPAD devices, the detection system generally selects a high-repetition-frequency detection system to obtain photon data quickly. Currently, random sequence coding is primarily used to suppress the range ambiguity effect caused by a high repetition frequency. To meet the requirements of fast target detection under the conditions of a long distance and large field of view, more complete target information must be obtained through multiple parallel SPAD detection pixels. For an SPAD array with a large array structure, any photon signal reaching the receiver surface can cause an equivalent photocurrent response. The return signal received by a certain element may originate from several different transmission signals, which may affect the array detection results. To effectively distinguish the array pixels at the receiver, the coding waveforms in each period should not interfere with each other, and numerous random coding sequences are required to be orthogonal to each other. Because of factors, such as waveform duration and hardware system accuracy, the coding sequences in the actual system have a certain degree of correlation. In addition, the optical signal is non-negative, and the power is only superimposed but not offset. The cross-correlation responses among the array elements accumulate, which interferes with the identification of correlation peaks, resulting in false detection alarms and result misjudgment. Therefore, identifying a random sequence with a strong anti-interference ability to encode an SPAD array is required for highly accurate detection.

The basic principle of using random coding to solve the range ambiguity problem in a single-photon detection system is introduced, and the coding structure and correlation characteristics of pseudo-random, true random, and chaotic sequences are compared and analyzed. Owing to the fixed coding length and form, pseudo-random sequences represented by m-sequences cannot generate codes of a given length according to the design requirements, and even longer coding sequences might need to be chosen. A long coding sequence inevitably leads to long encoding processing and receiving times, which has a significant impact on the delay in array detection. Although the true random sequence has a strong anti-crosstalk ability, its unpredictable and unrepeatable characteristics require a high-quality physical entropy source, which is limited in this application scenario. Chaotic sequences exhibit strong random and anti-interference characteristics. Their length is not limited by a fixed format and can be adjusted according to the greatest distance from the target, which can better meet the coding requirements of large-scale SPAD array detection. Through the simulation and analysis of the coding structure and correlation of various chaotic sequences, the application advantages of the improved logistic coding in large-scale spatiotemporal arrays were verified. Given the possible dynamic degradation of single-stage chaotic sequences, a compound logistic sequence is designed as the encoding method for the space-time array according to the range change of the Lyapunov exponent.In array-coded detection, the autocorrelation of sequences is an important index that affects the detection ability of a single array element, and the cross-correlation reflects the interference intensity between different detection elements. The Peak Side Lobe Ratio (PSLR) is typically used as an index to evaluate coding correlations within an array. However, the PSLR cannot be used to quantitatively compare the effect of cross-correlation accumulation on autocorrelation results in arrays. To quantify the effect of cross-correlation accumulation on the recognition of autocorrelation peaks in arrays, we established the concept of Peak Side-Lobe Difference (PSLD) to represent the normalized difference between the correlation peak and the maximum side-lobe value.

For encoding requirements of large-scale arrays, we set up a selection basis function based on PSLD, and use this function to filter the generated chaotic sequences to ensure that they can meet the anti-interference requirements of the desired array size. Because of the tediousness of calculating cross-correlations between arrays, a fast method for selecting available encoding sequences is provided by simulating and testing the changes in the autocorrelation PSLD and cross-correlation PSLD for multiple sets of sequences. Combined with the application requirements for long-range detection, a specific flow of large-scale SPAD array coding is provided.In order to meet the detection requirements of larger arrays, combined with the application requirements of long-range detection, the specific process of large-scale SPAD array encoding using traversal testing method is given. The codes used are filtered by PSLD, and each element can use a relatively fixed coding sequence, which makes the proposed scheme more advantageous in engineering implementation.

In order to meet the encoding requirements of large-scale SPAD arrays, through the comparative analysis of the number of encoding, generation efficiency, format length, correlation properties of various random sequences, this paper points out the advantages of improved Logistic encoding in large-scale SPAD arrays. Based on the variation of Lyapunov index range, a composite Logistic sequence is designed as the encoding method of large-scale SPAD arrays. At the same time, considering the requirement of no interference between the elements, the suitable relationship between the generated sequence and the size of the array and the optimal selection scheme are given. Next, we will focus on the design of appropriate chaos cascade based on the actual detection needs, and continue to explore better random coding schemes.